(19)
(11)EP 2 903 256 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
29.07.2020 Bulletin 2020/31

(21)Application number: 13842258.9

(22)Date of filing:  30.05.2013
(51)International Patent Classification (IPC): 
G06T 7/73(2017.01)
G06K 9/62(2006.01)
G06K 9/00(2006.01)
(86)International application number:
PCT/JP2013/065116
(87)International publication number:
WO 2014/050210 (03.04.2014 Gazette  2014/14)

(54)

IMAGE PROCESSING DEVICE, IMAGE PROCESSING METHOD AND PROGRAM

BILDVERARBEITUNGSVORRICHTUNG, BILDVERARBEITUNGSVERFAHREN UND PROGRAMM

DISPOSITIF DE TRAITEMENT D'IMAGE, PROCÉDÉ DE TRAITEMENT D'IMAGE ET PROGRAMME


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 26.09.2012 JP 2012213030

(43)Date of publication of application:
05.08.2015 Bulletin 2015/32

(73)Proprietor: Rakuten, Inc.
Tokyo 158-0094 (JP)

(72)Inventors:
  • HIRANO, Hiromi
    Tokyo 158-0094 (JP)
  • BANDARA, Udana
    Tokyo 158-0094 (JP)

(74)Representative: Hoffmann Eitle 
Patent- und Rechtsanwälte PartmbB Arabellastraße 30
81925 München
81925 München (DE)


(56)References cited: : 
JP-A- H01 239 406
JP-A- 2009 042 900
JP-A- 2012 124 745
US-A1- 2003 038 801
US-A1- 2004 227 751
US-A1- 2012 082 347
JP-A- 2003 271 045
JP-A- 2011 013 121
US-A1- 2001 014 171
US-A1- 2004 164 956
US-A1- 2005 195 157
  
  • John W. Roach et al.: "Determining the Movement of Objects from a Sequence of Images", IEEE Transactions on Pattern Analysis and Machine Intelligence, 6 November 1980 (1980-11-06), pages 554-562, XP055531356, Retrieved from the Internet: URL:https://ieeexplore.ieee.org/ielx5/34/6 447692/06447703.pdf?tp=&arnumber=6447703&i snumber=6447692 [retrieved on 2018-12-07]
 
Remarks:
The file contains technical information submitted after the application was filed and not included in this specification
 
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

Technical Field



[0001] The present invention relates to an image processing device, an image processing method, a program, and an information storage medium.

Background Art



[0002] As a method to shoot(image) an entire circumference of an object, a method is known in which an object is shot while mechanically rotating a turntable, on which the object is placed at the center, at a constant angular velocity (Non-Patent Literature 1). In addition, an application program is also known in which a user is allowed to rotate an object displayed on a screen arbitrarily with use of the captured image data obtained in this manner (Non-Patent Literature 1).

[0003] In the application program disclosed in Non-Patent Literature 1, it is configured that, when a user slides a finger on a touch panel, a frame of the captured image data moves back and forth by a number based on the slide amount. As a result, the object is rotated by an angle based on the slide amount of the user, and thus the user can view the object from any direction. Such a technology enables a user, for example, to view a product from an arbitrary direction on a website where the product is sold.

Citation List


Non-Patent Document



[0004] Non-Patent Literature 1: "The Coolest AppAtSXSW-Arqball Spin". [online]. March 13, 2012. [retrieved on July 26,2012]. Retrieved from the Internet: <URL:http://erickschonfeld.com/2012/03/13/coolest-app-sxsw-arqball-spin/>.

[0005] In US 2004/0227751 A1, a method of generating a three-dimensional image of an object includes placing a video camera at a predetermined distance from the object, such that the video camera has an unobstructed view of the object, and causing the object to rotate about a central axis. A video stream of at least one revolution of the rotating object is captured with the video camera. A period of the at least one revolution of the object is determined. A predetermined number of frames of the captured video stream are selected, and a three-dimensional image of the object using the selected frames is created.

[0006] US 2012/0082347 A1 describes an apparatus and a method for measuring a speed of a rotation body and recognizing a spin using a line scan. An apparatus and a method for calculating a motion of an object are capable of acquiring line scan images using some lines of an area scan camera and calculating a three-dimensional speed and a three-dimensional spin of a rotation body by using a composite image in which the line scan images are coupled. This can provide a realistic game or training at low cost by providing realistic physical simulation of the rotation body while allowing a competitive price when producing products by using the existing inexpensive camera.

[0007] In US 2003/0038801 A1, Images of a reference object for detecting each camera position are captured by cameras and the relative position between the cameras is obtained by using the captured images. Then, a target is placed on a turntable onto which reference marks for detecting position are formed, and while the turntable is rotated, images are captured by the cameras. Based on the reference marks on the turntable, the positional relationship between the camera and the target is determined, and, based on the relative position of the cameras, the positional relationship between the target and the camera is determined. A three-dimensional model of the target is produced based on the images captured by the cameras. In this manner, a target of relatively large size is easily three-dimensionally modeled.

Summary of Invention


Technical Problem



[0008] The above described technology requires a turntable that rotates mechanically at a constant angular velocity, and a preparation of such a turntable takes considerable costs or labors. In contrast, in the case of a turntable that is manually rotated, costs and labors can be reduced.

[0009] However, if a turntable that is rotated manually is used, the angular velocity of the object is not constant, and thus a rotation angle of the object for each frame is not constant in the captured image data. As such, when such captured image data is reproduced as it is, a rotational speed of the object is not constant. Further, in a case where such captured image data is used in the above described application program as it is, the rotation angle of the object based on the user's slide amount is not constant. As a result, sometimes users may feel uncomfortable.

[0010] In order to solve such an inconvenience, it is necessary to provide technology for obtaining moving image data in which a rotation angle of an object is approximately constant for each frame, based on captured image data obtained by shooting an object rotating at a non-constant angular velocity.

[0011] One or more embodiments of the present invention have been conceived in view of the above, and an object thereof is to provide an image processing device, an image processing method, a program, and an information storage medium capable of obtaining moving image data in which a rotation angle of an object is approximately constant for each frame, based on captured image data obtained by shooting an object rotating at a non-constant angular velocity.

Solution to Problem



[0012] The present invention provides an image processing device, as set out in claim 1, an image processing method, as set out in claim 14, and a program as set out in claim 15. Further embodiments are defined by the appended dependent claims.

[0013] In a case where the one captured image is selected as a frame image of the moving image data and the rotation angle information is equal to or more than the threshold value, the moving image data generating means may select, as another frame image of the moving image data, either one of the another captured image and a captured image that was shot immediately before the another captured image.

[0014] In a case where the rotation angle information relating to a rotation angle of the object between the point of time when the one captured image is shot and the point of time when the another captured image is shot is closer to the threshold value than rotation angle information relating to a rotation angle of the object between the point of time when the one captured image is shot and a point of time when the captured image that was shot immediately before the another captured image is shot, the moving image data generating means may select the another captured image as a frame image of the moving image data, and, in a case where the rotation angle information relating to the rotation angle of the object between the point of time when the one captured image is shot and the point of time when the captured image that was shot immediately before the another captured image is shot is closer to the threshold value than the rotation angle information relating to the rotation angle of the object between the point of time when the one captured image is shot and the point of time when the another captured image is shot, the moving image data generating means may select the captured image that was shot immediately before the another captured image as a frame image of the moving image data.

[0015] In a case where a feature point corresponding to a certain point on the object is specified on a first pair of captured images, the position information obtaining means may obtain position information relating to a position of the feature point in one captured image that is shot first among the first pair of captured images and in another captured image that is shot later among the first pair of captured images. The rotation angle information obtaining means may obtain rotation angle information relating to a rotation angle of the object between a point of time when the one captured image of the first pair of captured images is shot and a point of time when the another captured image of the first pair of captured images is shot based on the position information of the first pair of captured images. In a case where a feature point corresponding to a certain point on the object is specified in a second pair of captured images, the position information obtaining means may obtain position information relating to a position of the feature point of one captured image that is shot first among the second pair of captured images and another captured image that is shot later among the second pair of captured images. The rotation angle information obtaining means may obtain rotation angle information relating to a rotation angle of the object between a point of time when the one captured image of the second pair of captured images is shot and a point of time when the another captured image of the second pair of captured images is shot based on the position information of the second pair of captured images. In a case where the another captured image of the first pair of captured images is identical to the one captured image of the second pair of captured images, the moving image data generating means may include means for obtaining cumulative rotation angle information relating to a rotation angle of the object between the point of time when the one captured image of the first pair of captured images is shot and the point of time when the another captured image of the second pair of captured images is shot, based on the rotation angle information of the first pair of captured images and the rotation angle information of the second pair of captured images. The moving image data generating means may select a plurality of frame images of the moving image data from the plurality of captured images based on the cumulative rotation angle information.

[0016] In a case where the one captured image of the first pair of captured images is selected as a frame image of the moving image data, the moving image data generating means may determine whether or not to select the another captured image of the second pair of captured images as another frame image of the moving image data, based on a result of comparing the cumulative rotation angle information with the threshold value.

[0017] In a case where the one captured image of the first pair of captured images is selected as a frame image of the moving image data and the cumulative rotation angle information is equal to or more than the threshold value, the moving image data generating means may select either one of the another captured image of the second pair of captured images and a captured image that was shot immediately before the another captured image of the second pair of captured images as another frame image of the moving image data.

[0018] The rotation angle information obtaining means may include means for obtaining position difference information relating to a difference between the position of the feature point in the one captured image and the position of the feature point in the another captured image, and means for obtaining the rotation angle information based on the position difference information.

[0019] The image processing device may include means for setting an area in the captured image as an observation area based on a position of a rotation axis of the object in the captured image, and the rotation angle information obtaining means may obtain the rotation angle information based on the position difference information in a case where the feature point is included in the observation area.

[0020] The rotation angle information obtaining means may obtain the rotation angle information using the position difference information of a feature point whose rotation radius of when the object rotates is large, in preference to the position difference information of a feature point whose rotation radius of when the object rotates is small.

[0021] The rotation angle information obtaining means may include rotation radius information obtaining means for obtaining, in a case where a plurality of feature points respectively corresponding to a plurality of points on the object are specified in the one captured image and the another captured image, rotation radius information relating to a rotation radius of a point on the object corresponding to the feature point, for each of the plurality of feature points, means for obtaining, for each of the plurality of feature points, rotation angle information relating to a rotation angle of the point corresponding to the feature point on the object between a point of time when the one captured image is shot and a point of time when the another captured image is shot, based on the position difference information and the rotation radius information of the feature points, and means for obtaining the rotation angle information of the object based on the rotation angle information of each of the plurality of feature points.

[0022] The rotation radius information obtaining means may include means for obtaining, in a case where the feature point corresponding to the certain point on the object is specified in at least three captured images, elliptical orbit information relating to an elliptical orbit of the feature point based on the position of the feature point in the at least three captured images, and means for obtaining the rotation radius information of the feature point based on the elliptical orbit information of the feature point.

[0023] In a case where the feature point corresponding to the certain point on the object is first specified in an ith (i: an integer of 1 or more) captured image and specified last in a jth (j: an integer greater than i) captured image that is shot later than the ith captured image, the rotation radius information obtaining means may obtain the rotation radius information of the feature point based on a position of the feature point in the ith captured image and a position of the feature point in the jth captured image.

[0024] The rotation angle information obtaining means may include means for obtaining, in a case where the feature point corresponding to the certain point on the object is specified in at least three captured images, elliptical orbit information relating to an elliptical orbit of the feature point and means for obtaining the rotation angle information based on the positions of the feature point in the one captured image and in the another captured image and the elliptical orbit information.

[0025] The moving image data generating means may include means for obtaining similarity information relating to a similarity between a reference captured image that is selected from the plurality of captured images and each of a plurality of captured images that are shot after a predetermined period of time passes from a point of time when the reference captured image is shot, and means for selecting one of the plurality of captured images that are shot after the predetermined period of time passes from the point of time when the reference captured image is shot as a captured image of a point of time when the object completes one rotation based on the similarity information. The moving image data generating means may generate the moving image data showing how the object rotates once based on captured images that are shot before the point of time when the captured image selected as the captured image of the time when the object completes one rotation is shot.

[0026] The moving image data generating means may include feature point number obtaining means for obtaining a number of feature points corresponding to feature points specified in the reference captured image, from among the feature points specified in the captured image that is shot after the predetermined period of time passes from the point of time when the reference captured image is shot, means for selecting candidates of the captured image of the point of time when the object completes one rotation, from among the plurality of captured images that are shot after the predetermined period of time passes from the point of time when the reference captured image is shot, based on a result obtained by the feature point number obtaining means, means for obtaining similarity information relating to a similarity between the reference captured image and each of the captured images selected as the candidates, and means for selecting one of the captured images selected as the candidates as the captured image of a point of time when the object completes one rotation based on the similarity information.

[0027] The plurality of captured images may include the 1st to mth (m: an integer of 2 or more) captured images and m+1th to m+nth (n: an integer of 1 or more) captured images, which are shot after the mth captured image. The moving image data generating means may include selection number obtaining means for obtaining a number of the captured images selected as frame images of the moving image data out of the 1st to the mth captured images, and means for selecting the frame images of the moving image data out of the m+1th to the m+nth captured images, in a case where the feature point corresponding to the point on the object is not specified in the m+1th to the m+nth captured images, based on a ratio of a number of the 1st to the mth captured images to the number of the captured images obtained by the selection number obtaining means.

[0028] The plurality of captured images may include the 1st to mth (m: an integer of 2 or more) captured images and m+1th to m+nth (n: an integer of 1 or more) captured images, which are shot after the mth captured image. The moving image data generating means may include selection number obtaining means for obtaining a number of the captured images selected out of the m+1th to the m+nth captured images as the frame images of the moving image data, and means for selecting the frame images of the moving image data out of the 1st to the mth captured images, in a case where the feature point corresponding to the point on the object is not specified in the 1st to the mth captured images, based on a ratio of a number of the m+1th to the m+nth captured images to the number of the captured images obtained by the selection number obtaining means.

[0029] The image processing device may include first determination means for determining whether or not a feature amount of the feature point specified in the one captured image and a feature amount of the feature point specified in the another captured image area are the same or substantially the same, second determination means for determining whether or not a difference of a coordinate value of a direction of an axis corresponding to a rotation axis of the object between the position of the feature point specified in the one captured image and the position of the feature point specified in the another captured image is equal to or more than the threshold value, and means for determining whether or not the feature point specified in the one captured image and the feature point specified in the another captured image are feature points corresponding to the certain point on the object based on a determination result of the first determination means and the second determination means.

[0030] The image processing device may include determination means for determining whether or not the feature amount of the feature point specified in the one captured image and the feature amount of the feature point specified in the another captured image are the same or substantially the same, comparing means for comparing (a) information relating to a difference of a coordinate value of a direction of a second axis perpendicular to a first axis corresponding to a rotation axis of the object between the position of one feature point in the one captured image and the position of the one feature point in the another captured image with (b) information relating to a difference of a coordinate value of the direction of the second axis between the position of another feature point in the one captured image and the position of the another feature point in the another captured image, in a case where the one feature point having the same or substantially the same feature amount in both of the one captured image and the another captured image, and the another feature point having the same or substantially the same feature amount in both of the one captured image and the another captured image are specified, and means for determining whether or not the feature point specified in the one captured image and the feature point specified in the another captured image are feature points corresponding to the certain point on the object, based on a determination result of the determination means and a comparing result of the comparing means.

[0031] The captured image obtaining means may obtain a plurality of captured images that are obtained by repeatedly shooting the object while the object rotates a plurality of times. The moving image data generating means may generate the moving image data that shows how the object rotates once based on the plurality of captured images that are obtained by repeatedly shooting the object while the object rotates the plurality of times.

Advantageous Effects of Invention



[0032] According to the present invention, it is possible to obtain moving image data in which a rotation angle of an object is approximately constant for each frame based on captured image data of the object rotating at a non-constant angular velocity.

Brief Description of Drawings



[0033] 

[FIG. 1] A diagram illustrating an example of a hardware configuration of an image processing device according to an embodiment of the present invention.

[FIG. 2] A diagram for explaining a method of shooting an item.

[FIG. 3] A diagram for explaining a method of shooting an item.

[FIG. 4] A diagram for explaining an outline of a method for generating moving image data of an item.

[FIG. 5] A functional block diagram illustrating functions implemented in the image processing device.

[FIG. 6] A diagram illustrating an example of a feature point specified by a feature point specifying unit.

[FIG. 7] A diagram illustrating an example of an orbit of a feature point corresponding to the certain point on the item.

[FIG. 8] A diagram illustrating an example of feature points specified in one captured image and another captured image.

[FIG. 9] A diagram for explaining a method for obtaining rotation angle information.

[FIG. 10] A diagram illustrating the Sin function.

[FIG. 11] A diagram for explaining a method for setting an observation area.

[FIG. 12] A diagram for explaining a method for setting an observation area.

[FIG. 13] A diagram for explaining an example of a method for selecting frame images.

[FIG. 14] A diagram for explaining another example of a method for selecting frame images.

[FIG. 15] A diagram for explaining the other example of a method for selecting frame images.

[FIG. 16] A flow chart showing an example of processing executed in the image processing device.

[FIG. 17] A flow chart showing an example of processing executed in the image processing device.

[FIG. 18] A diagram for explaining an outline of processing executed in the image processing device.

[FIG. 19] A flow chart showing details of processing executed in the image processing device.

[FIG. 20] A flow chart showing another example of processing executed in the image processing device.

[FIG. 21] A flow chart showing another example of processing executed in the image processing device.

[FIG. 22A] A diagram illustrating a variation.

[FIG. 22B] A diagram illustrating a variation.

[FIG. 23] A diagram illustrating another variation.


Description of Embodiments



[0034] Examples of an embodiment of the present invention will be described below in detail with reference to the accompanying drawings.

[0035] FIG. 1 illustrates an example of a hardware configuration of an image processing device according to an embodiment of the present invention. The image processing device 10 according to this embodiment is implemented by, for example, a personal computer, a server computer, a mobile phone (including a smart phone), or a portable information terminal (including a tablet computer).

[0036] As shown in FIG. 1, the image processing device 10 includes a control unit 11, a storage unit 12, an optical disc drive unit 13, a communication unit 14, an operation unit 15, a display unit 16, and an audio output unit 17.

[0037] The control unit 11 includes, for example, one or more microprocessors, and executes processing according to an operating system or a program stored in the storage unit 12. The storage unit 12 includes a main storage unit, such as a RAM, and an auxiliary storage unit, such as a hard disk or a solid-state drive.

[0038] The optical disc drive unit 13 reads a program and data stored in an optical disc (information storage medium) . The program and data are provided to the storage unit 12 through the optical disc. That is, the program and data stored in the optical disc are read and stored in the storage unit 12. The optical disc drive unit 13 is not an essential element. The image processing device 10 may include an element for reading the program and data stored in an information storage medium other than the optical disc (e.g., memory card). The program and data may be provided to the storage unit 12 through an information storage medium other than the optical disc.

[0039] The communication unit 14 is an interface for connecting the image processing device 10 to a communication network. The image processing device 10 is configured to send and receive data to and from other devices through the communication network. The program and data may be provided to the storage unit 12 through the communication network.

[0040] The operation unit 15 is a unit for a user to operate. The operation unit 15 includes, for example, a keyboard, a pointing device or the like. The display unit 16 is, for example, a liquid crystal display, an organic EL display or the like, and the audio output unit 17 is, for example, speakers, headphones or the like.

[0041] The following explains technology of generating moving image data in which a rotation angle of an object is approximately constant for each frame, based on captured image data of an object rotating at a non-constant angular velocity. In the following, a case is discussed in which moving image data of commercial items is generated. However, the present invention is applicable to a case where moving image data of an object other than a commercial item is generated.

[0042] First, a method for shooting an item will be discussed. FIGs. 2 and 3 illustrate a method for shooting an item. As shown in FIG. 2, when an item 30 (beckoning cat in the example of FIG. 2) is shot, the item 30 is placed on a turntable 20.

[0043] The turntable 20 includes a thick board 22, a cardboard 24, and a thumbtack 28. The thick board 22 is a board in which the thumbtack 28 can be safely stuck, and the cardboard 24 is attached to the thick board 22 with the thumbtack 28. The thumbtack 28 is stuck in the center 26 of the cardboard 24.

[0044] The item 30 is placed at the center of the cardboard 24. The item 30 is placed so that the center axis of the item 30 matches the center 26 of the cardboard. The cardboard 24 is rotatable with the thumbtack 28 as a rotation axis 32. When a videographer rotates the cardboard 24 manually, the item 30 placed on the cardboard 24 rotates about the rotation axis 32.

[0045] As shown in FIG. 3, an shooting device 34 for shooting the item 30 is fixed at the position from which the item 30 placed on the turntable 20 can be shot. In the example shown in FIG. 3, the shooting device 34 is placed in front of the turntable 20 and slightly above the item 30.

[0046] The shooting device 34 includes a digital video camera, and can shoot a still image and a moving image. For example, a single digital video camera or a digital video camera built in a mobile phone etc. is used as the shooting device 34. Further, for example, a digital video camera built in the image processing device 10 may be used as the shooting device 34. A digital camera capable of shooting only still images may be used as the shooting device 34.

[0047] The shooting device 34 performs shooting at predetermined intervals (e.g., every 1/30 seconds) . In other words, the shooting device 34 shoots a moving image having the predetermined number (e.g., 30) of frames per second.

[0048] The videographer starts shooting with the shooting device 34, and also rotates the cardboard 24, on which the item 30 is placed, counterclockwise (or clockwise). For example, the videographer rotates the cardboard 24 using an index finger and a middle finger alternately. In this case, a moving image of the rotating item 30 is shot by the shooting device 34. This moving image is stored in the storage unit 12 of the image processing device 10 through a storage medium or a communication network. This moving image may be stored in a device other than the image processing device 10.

[0049] Instead of shooting a moving image of the item 30, the videographer may continuously shoot still images of the item 30 while the item 30 is rotating. A group of the still images obtained as described above may be used in place of the above moving image.

[0050] In the following, a moving image or a group of still images shot by the shooting device 34 is referred to as "captured image data." Each image included in the captured image data is referred to as "captured image." For example, if a moving image shot by the shooting device 34 is captured image data, each frame image of the moving image corresponds to a "captured image." Further, for example, if a group of still images shot by the shooting device 34 is captured image data, each still image corresponds to a "captured image."

[0051] In the following, moving image data of an item 30 to be generated based on captured image data, that is, moving image data in which a rotation angle of an item 30 is approximately constant for each frame is simply referred to as "moving image data."

[0052] Next, a method for generating moving image data in which a rotation angle of an item 30 is approximately constant for each frame, based on the captured image data as above will be discussed below. FIG. 4 illustrates an outline of a method for generating moving image data.

[0053] In FIG. 4, the numbers attached above the captured images 40 included in the captured image data are image numbers. Each image number indicates a shooting order for a captured image. If the captured image data is a moving image, an image number corresponds to a frame number of the moving image. On the other hand, if the captured image data is a group of still images, an image number indicates where a still image in question is in a chronological shooting order. In the following, a captured image 40 with an image number "i", that is, the ith captured image 40 is written as a captured image [i].

[0054] In FIG. 4, the numbers attached above the frame images 42 included in the moving image data are frame numbers. In the following, a frame image 42 with a frame number "j", that is, the jth frame image 42 is written as the frame image [j].

[0055] As shown in FIG. 4, the image processing device 10 generates moving image data by selecting each frame image 42 constituting the moving image data from the captured images 40.

[0056] In particular, the image processing device 10 selects each frame images 42 so that a rotation angle (Δθ) of the item 30 in adjacent frame images 42 is approximately constant.

[0057] In other words, each frame image 42 is selected so that an angle (Δθ) between a representative direction (e.g., front direction) of the item 30 in a frame image 42 and a representative direction of the item 30 in a frame image 42 subsequent to the frame image 42 is approximately constant.

[0058] In still other words, each frame image 42 is selected so that a rotation angle (Δθ) of the item 30 between a point of time when a captured image 40 selected as a frame image 42 is shot and a point of time when a captured image 40 selected as a frame image 42 subsequent to the frame image 42 is shot is approximately constant.

[0059] In the example shown in FIG. 4, the captured image [1] is selected as the frame image [1]. The captured image [4] is selected as the frame image [2]. The captured image [6] is selected as the frame image [3].

[0060] Next, a constitution for implementing the generating method described above will be discussed. FIG. 5 is a functional block diagram illustrating functions implemented in the image processing device 10.

[0061] As shown in FIG. 5, the image processing device 10 includes a captured image obtaining unit 50, a feature point specifying unit 52, a position information obtaining unit 54, a rotation angle information obtaining unit 56, and a moving image data generating unit 58. These functional blocks are implemented by, for example, the control unit 11. That is, the control unit 11 executes processing according to a program, to thereby function as these functional blocks.

[0062] The captured image obtaining unit 50 will be discussed. The captured image obtaining unit 50 obtains a plurality of captured images 40 that are obtained by repeatedly shooting an object while the object is rotating. In the above described example, the item 30 corresponds to the "object."

[0063] That is, the captured image obtaining unit 50 reads captured image data stored in the storage unit 12. The captured image data may be stored in a device other than the image processing device 10. In other words, the captured image obtaining unit 50 may obtain captured image data stored in a device other than the image processing device 10 through a communication network.

[0064] The feature point specifying unit 52 and the position information obtaining unit 54 will be discussed. The feature point specifying unit 52 specifies feature points in captured images 40. In a case where a feature point corresponding to the certain point on the object is specified in a pair of captured images 40, the position information obtaining unit 54 obtains position information of a position of the feature point in one captured image 40 of the pair, which is shot earlier, and a position of the feature point in the other captured image 40 of the pair, which is shot later.

[0065] For example, the feature point specifying unit 52 obtains positions and feature amounts of the feature points in the captured images 40 using known algorithm for specifying a feature point. For example, the SURF (Speed-Up Robust Features) algorithm can be used as such an algorithm. An algorithm other than the SURF algorithm may also be utilized. For example, the SIFT (Scale-Invariant Feature Transform) algorithm and the Brisk (Binary Robust Invariant Scalable Keypoints) algorithm may also be utilized.

[0066] As will be discussed below, a feature point is used for obtaining information on a rotation angle of the item 30. As such, a feature point included in an area (hereinafter "background area") other than the area in which the item 30 is included is not necessary. Thus, the feature point specifying unit 52 may ignore a feature point in the background area.

[0067] For example, the feature point specifying unit 52 obtains a subtraction image between two consecutive captured images 40. The captured image 40 is a captured image of a rotating item 30, and thus, a background area of the captured image 40 is not basically changed. As such, the feature point specifying unit 52 assumes an area having zero pixel value in the subtraction image as a background area, and removes the feature points in this area. In other words, the feature points included in this area are not considered in the position information obtaining unit 54 and the rotation angle information obtaining unit 56.

[0068] FIG. 6 illustrates an example of a feature point specified by the feature point specifying unit 52. In FIG. 6, a feature point 60 corresponding to a nose of the beckoning cat is specified. In reality, a plurality of feature points 60 are specified in one captured image 40, but they are omitted in FIG. 6. The position of the feature point 60 is represented, for example, by coordinate values of an X-Y coordinate system, in which the upper left vertex of the captured image 40 is the origin O, and the rightward direction and the downward direction of the captured image 40 are the positive X-axis direction and the positive Y-axis direction, respectively.

[0069] For example, the feature point 60 corresponding to the nose of the beckoning cat is specified in each of the captured images 40 in which the nose of the beckoning cat is presented. In other words, the feature point 60 corresponding to the certain point on the item 30 is specified in the plurality of captured images 40.

[0070] FIG. 7 illustrates an example of an orbit of the feature point 60 corresponding to a certain point on the item 30. For example, FIG. 7 shows an orbit 70 of the feature point 60 corresponding to the nose of the beckoning cat. In the example shown in FIG. 3, the shooting device 34 is placed in front of the turntable 20 and slightly above the item 30, and thus, the orbit 70 of the feature point 60 is an elliptical orbit as shown in FIG. 7.

[0071] Basically, the feature amount of the feature point 60 corresponding to the nose of the beckoning cat is the same in each of the captured images 40. In other words, the feature points 60 corresponding to the certain point on the item 30 may be specified in the plurality of captured images 40, but basically, the feature amounts of the feature points 60 corresponding to the certain point on the item 30 is the same in each of the plurality of captured image 40.

[0072] Then the feature point specifying unit 52 (first determination means) determines whether or not a feature amount of a feature point 60 specified in a first captured image 40 is the same or substantially the same as a feature amount of a feature point 60 specified in a second captured image 40. When a difference between the feature amounts of these feature points 60 is equal to or less than a threshold value, the feature point specifying unit 52 determines the feature amounts of these feature points 60 are substantially the same. When the feature amounts of these feature points 60 are the same or substantially the same, the feature point specifying unit 52 determines that these feature points 60 are feature points 60 corresponding to a certain point on the item 30.

[0073] The item 30 rotates about the rotation axis 32 and the Y-axis is an axis corresponding to the rotation axis 32, and thus, the Y-axis coordinate values of the feature points 60 corresponding to the certain point on the item 30 do not differ from each other so much.

[0074] As such, the feature point specifying unit 52 (second determination means) determines whether or not a difference between a Y-axis coordinate value of a feature point 60 specified in a first captured image 40 and a Y-axis coordinate value of a feature point 60 specified in a second captured image 40 is a threshold value or more. If the difference between the Y-axis coordinate values of these feature points 60 is the threshold value or more, the feature point specifying unit 52 determines that these feature points 60 are not feature points 60 corresponding to a certain point on the item 30, even though the feature amounts of these feature points 60 are the same or substantially the same.

[0075] Further, when a plurality of feature points 60 respectively corresponding to certain points on the item 30 are specified in a first captured image 40 and a second captured image 40, change amounts of X-axis coordinate values of each of the plurality of feature points 60 are substantially the same, and are not significantly different from one another.

[0076] As such, as shown in FIG. 8, for example, in a case where a feature point 60A having the same or substantially the same feature amount in both of a first captured image 40 and a second captured image 40, a feature point 60B having the same or substantially the same feature amount in both of the first captured image 40 and the second captured image 40, a feature point 60C having the same or substantially the same feature amount in both of the first captured image 40 and the second captured image 40, and a feature point 60D having the same or substantially the same feature amount in both of the first captured image 40 and the second captured image 40 are specified, the feature point specifying unit 52 (comparing means) compares a change amount of the X-axis coordinate value of the feature point 60D with change amounts of X-axis coordinate values of other feature points 60A-60C.

[0077] In the example shown in FIG. 8, the change amounts of the X-axis coordinate values of the feature points 60A-60C are substantially the same, and the change amount of the X-axis coordinate value of the feature point 60D is greater than those of feature points 60A-60C. As described above, when the change amount of the X-axis coordinate value of the feature point 60D is largely different from the change amounts of the X-axis coordinate values of the feature points 60A-60C like this, the feature point specifying unit 52 determines that the feature point 60D is not a feature point 60 corresponding to the certain point in the item 30. In other words, the feature point specifying unit 52 determines that the feature point 60D, which has an absolute value of deviation of the change amount of the X-axis coordinate value greater than the threshold value, is not a feature point 60 corresponding to the certain point in the item 30.

[0078] The rotation angle information obtaining unit 56 will be discussed. The rotation angle information obtaining unit 56 obtains rotation angle information on a rotation angle of an object between the point of time when a first captured image 40 is shot and the point of time when a second captured image 40 is shot, based on the position information obtained by the position information obtaining unit 54.

[0079] "Rotation angle information" may be, for example, numerical information indicating a rotation angle itself, or numerical information which does not indicate a rotation angle itself but which is in proportion to a rotation angle. In other words, "rotation angle information" may be numerical information directly indicating a size of a rotation angle, or numerical information indirectly indicating a size of a rotation angle.

[0080] FIG. 9 is a diagram for explaining a method for obtaining rotation angle information and a relationship between a rotation angle of the item 30 and changes in X-coordinate values of a feature point 60. FIG. 9 illustrates the item 30 placed on the turntable 20, viewed from right above.

[0081] When the item 30 rotates around the rotation axis 32, an orbit 80 of a point 30A on the item 30 is a circular orbit centered on the rotation axis 32. In the following, the radius of the circular orbit is written as "r."

[0082] For example, in a case where the item 30 rotates only by a rotation angle θ1, a change amount x1 of the X-axis coordinate value of the feature point 60 corresponding to the point 30A on the item 30 is represented by a below equation (1). That is, x1 is a value obtained by multiplying r by sinθ1. Similarly, in a case where the item 30 rotates only by a rotation angle θ2, a change amount x2 of the X-axis coordinate value of the feature point 60 corresponding to the point 30A on the item 30 is represented by a below equation (2). That is, x2 is a value obtained by multiplying r by sinθ2.





[0083] The Sin function is as what is shown in FIG. 10, and then the Sin function can be linearly approximated at a nearby area 90 where the angle θ is zero or close to zero. As such, the above equations (1) and (2) can be approximated as below equations (3) and (4) using a proportional constant a. Further, based on equations (3) and (4), a relationship between x1, x2, θ1, and θ2 is represented by a below equation (5).







[0084] Further, x2 and θ2 can be represented by below equations (6) and (7).





[0085] The above equation (5) can be modified as below equation (8) . The below equation (8) can be arranged as below equation (9). Further, based on the above equation (3), the below equation (9) can be modified as below equation (10). Then, below equation (11) can be obtained based on the below equation (10).









[0086] The above equation (11) indicates that a change of the angle (Δθ) is proportional to a change of the X-axis coordinate value (Δx). Accordingly, the rotation angle (Δθ) of when the item 30 rotates has a proportional relation with the change amount (Δx) of the X-axis coordinate value of the feature point 60 corresponding to the point 30A on the item 30.

[0087] Now, a radius r in the equation (11) basically varies at each point on the item 30. However, in a case where the item 30 has a cylindrical shape, the above radius r does not vary at each point on the item 30 and substantially the same. As such, assuming that a shape of the item 30 is cylindrical, the radius r in the equation (11) may be ignored, and thus, the change amount of the X-axis coordinate value of the feature point 60 may be used as a value indicating a size of a rotation angle of the item 30.

[0088] In view of the above, functions of the rotation angle information obtaining unit 56 will be discussed in detail.

[0089] The rotation angle information obtaining unit 56 sets an observation area in a captured image 40 based on the position of the rotation axis 32 in the captured image 40. The observation area is an area corresponding to the nearby area 90 shown in FIG. 10. FIGs. 11 and 12 are diagrams for explaining a method for setting the observation area.

[0090] Firstly, the rotation angle information obtaining unit 56 extracts a feature point 60 having the minimum X-axis coordinate value and a feature point 60 having the maximum X-axis coordinate value from among feature points 60 in all of the captured images 40. In this case, a feature point 60 included in the background area is ignored.

[0091] In FIG. 11, the X-axis coordinate value of the feature point 60 having the minimum X-axis coordinate value is written as "Xmin", and the X-axis coordinate value of the feature point 60 having the maximum X-axis coordinate value is written as "Xmax." The rotation angle information obtaining unit 56 assumes that an area where a X-axis coordinate value satisfies Xmin≦x≦Xmax as an area 100 in which a feature point 60 can exist (hereinafter referred to as an "existence area").

[0092] Subsequently, the rotation angle information obtaining unit 56 obtains a position of the rotation axis 32 in the captured image 40 based on the existence area 100. As shown in FIG. 11, for example, the rotation angle information obtaining unit 56 assumes a straight line 102 that equally divides the existence area 100 toward right and left as a straight line corresponding to the rotation axis 32. In other words, the rotation angle information obtaining unit 56 assumes the straight line 102, which passes through the intermediate value of the Xmin and the Xmax, ((Xmin + Xmax)/2), and is parallel to the Y-axis, as a straight line corresponding to the rotation axis 32.

[0093] The rotation angle information obtaining unit 56 may obtain an object area, in which an object (item 30) is presented, in each of all the captured images 40. Then, the rotation angle information obtaining unit 56 may assume a straight line that equally divides a sum-set area of those object areas toward right and left as a straight line corresponding to the rotation axis 32.

[0094] The rotation angle information obtaining unit 56 sets an observation area based on the straight line 102 obtained as described above. As shown in FIG. 12, for example, the rotation angle information obtaining unit 56 sets a rectangular area with the straight line 102 passing through the center and having the width d as the observation area 110. In this case, the width d is set as a value smaller than a width w of the existence area 100.

[0095] As described later, the rotation angle information obtaining unit 56 uses only a feature point 60 included in the observation area 110 in order to obtain rotation angle information. For this reason, if a value of the width d is too small, a number of the feature points 60 included in the observation area 110 gets small, and thus an accuracy of the rotation angle information may possibly be reduced. On the other hand, if a value of the width d is too big, the observation area 110 may not possibly correspond to the nearby area 90 in FIG. 10. In this case, the premise that a linear approximation of the Sin function is possible is not satisfied, and as a result, the accuracy of the rotation angle information may possibly be reduced. It is preferable, to set the value of width d as, for example, about one third of the value of width (w) of the existence area 100.

[0096] In a case where the feature point 60 corresponding to the certain point on the item 30 is specified in the observation areas 110 in both of a first captured image 40 and a second captured image 40, the rotation angle information obtaining unit 56 obtains position difference information regarding a difference between a position of a feature point 60 in the first captured image 40 and a position of a feature point 60 in the second captured image 40.

[0097] For example, the rotation angle information obtaining unit 56 obtains information regarding differences in X-axis coordinate values as the above position difference information. As described above, the X-axis is an axis perpendicular to the Y-axis corresponding to the rotation axis 32.

[0098] The rotation angle information obtaining unit 56 obtains rotation angle information based on the above position difference information.

[0099] For example, when a plurality of feature points 60 respectively corresponding to certain points on the item 30 are specified in the observation areas 110 in both of a first captured image 40 and a second captured image 40, the rotation angle information obtaining unit 56 calculates a change amount of X-axis coordinate values between the first captured image 40 and the second captured image 40 for each of the feature points 60. Subsequently, the rotation angle information obtaining unit 56 calculates statistics (e.g., average value) of the change amounts of the respective feature points 60. In this case, such calculated values are used as the rotation angle information.

[0100] Alternatively, the rotation angle information obtaining unit 56 may calculate center of gravity positions of the above feature points 60 in the first captured image 40. Further, the rotation angle information obtaining unit 56 may calculate center of gravity positions of the above feature points 60 in the second captured image 40. Then, the rotation angle information obtaining unit 56 may calculate differences in X-axis coordinate values between those center of gravity positions. In this case, such calculated values are used as the rotation angle information.

[0101] The values described above correspond to numerical values representing sizes of rotation angles, but are not numerical values representing the rotation angles themselves. For this reason, in the following, such numerical values are referred to as "pseudo-rotation angle" for convenience.

[0102] The moving image data generating unit 58 will be discussed. The moving image data generating unit 58 generates moving image data by selecting, from among the captured images 40, each frame images 42 of the moving image data showing how the object rotates based on the position information obtained by the position information obtaining unit 54.

[0103] For example, the moving image data generating unit 58 selects, from among the captured images 40, frame images 42 of the moving image data showing how the item 30 rotates based on the rotation angle information obtained by the rotation angle information obtaining unit 56 for each of the captured images 40 included in the captured image data.

[0104] Specifically, the moving image data generating unit 58 selects each frame images 42 so that rotation angles of the item 30 between the frame images 42 are substantially the same.

[0105] FIG. 13 is a diagram for explaining an example of a method for selecting frame images 42. In FIG. 13, for example, Δθ1,2 represents a pseudo-rotation angle between a captured image [1] and a captured image [2]. That is, Δθ1,2 represents a pseudo-rotation angle of the item 30 between the point of time when the captured image [1] is shot and the point of time when the captured image [2] is shot. For example, in a case where a plurality of feature points 60 respectively corresponding to certain points on the item 30 are specified in the observation areas 110 in both of the captured image [1] and the captured image [2], an average value of the change amounts of the X-axis coordinate values of the feature points 60 between the captured image [1] and the captured image [2] corresponds to Δθ1,2.

[0106] Firstly, the moving image data generating unit 58 selects a frame image [1]. For example, the moving image data generating unit 58 selects the captured image [1] as the frame image [1].

[0107] Subsequently, the moving image data generating unit 58 selects a frame image [2]. The moving image data generating unit 58 first determines whether or not to select the captured image [2], which follows the captured image [1] selected as the preceding frame image [1], as the frame image [2].

[0108] In this case, the moving image data generating unit 58 compares a pseudo-rotation angle Δθ1,2 between the captured image [1] and the captured image [2] with a threshold value θT. In other words, the moving image data generating unit 58 determines whether the pseudo-rotation angle Δθ1,2 between the captured image [1] and the captured image [2] is equal to or more than the threshold value θT. The threshold value θT is a proper pseudo-rotation angle as a pseudo-rotation angle between the frame images 42, and is determined in advance.

[0109] If the pseudo-rotation angle Δθ1,2 is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [2] as the frame image [2]. On the other hand, as shown in FIG. 13, if the pseudo-rotation angle Δθ1,2 is less than the threshold value θT, the moving image data generating unit 58 does not select the captured image [2] as the frame image [2] .

[0110] If the captured image [2] is not selected as the frame image [2], the moving image data generating unit 58 determines whether or not to select the following captured image [3] as the frame image [2] .

[0111] In this case, the moving image data generating unit 58 compares a cumulative pseudo-rotation angle between the captured image [1] selected as the preceding frame image [1] and the captured image [3] with the threshold value θT. That is, the moving image data generating unit 58 determines whether the sum of the pseudo-rotation angle Δθ1,2 between the captured images [1] and [2] and the pseudo-rotation angle Δθ2,3 between the captured images [2] and [3] is equal to or more than the threshold value θT. In this case, the sum corresponds to a numerical value indicating a size of the rotation angle of the item 30 between the point of time when the captured image [1] is shot and the point of time when the captured image [3] is shot.

[0112] If the sum of the pseudo-rotation angles Δθ1,2 and Δθ2,3 is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [3] as the frame image [2]. On the other hand, as shown in FIG. 13, the sum of the pseudo-rotation angles Δθ1,2 and Δθ2,3 is less than the threshold value θT, the moving image data generating unit 58 does not select the captured image [3] as the frame image[2].

[0113] In a case where the captured image [3] is not selected as the frame image [2], the moving image data generating unit 58 determines whether or not to select the following captured image [4] as the frame image [2].

[0114] In this case, the moving image data generating unit 58 determines whether the cumulative pseudo-rotation angle between the captured image [1] selected as the preceding frame image [1] and the captured image [4] is equal to or more than the threshold value θT. That is, the moving image data generating unit 58 determines whether the sum of the pseudo-rotation angle Δθ1,2 between the captured images [1] and [2], the pseudo-rotation angle Δθ2,3 between the captured images [2] and [3], and the pseudo-rotation angle Δθ3,4 between the captured images [3] and [4] is equal to or more than the threshold value θT. In this case, the sum corresponds to a numerical value indicating a size of the rotation angle of the item 30 between the point of time when the captured image [1] is shot and the point of time when the captured image [4] is shot.

[0115] As shown in FIG. 13, if the sum of the pseudo-rotation angles Δθ1,2, Δθ2,3, and Δθ3,4 is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [4] as the frame image [2].

[0116] After selecting the frame image [2], the moving image data generating unit 58 selects the frame [3]. The moving image data generating unit 58 first determines whether or not to select the captured image [5], which follows the captured image [4] selected as the preceding frame image [2], as the frame image [3].

[0117] In this case, the moving image data generating unit 58 determines whether the pseudo-rotation angle Δθ4,5 between the captured image [4] and the captured image [5] is equal to or more than the threshold value θT. If the pseudo-rotation angle Δθ4,5 is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [5] as the frame image [3]. On the other hand, as shown in FIG. 13, if the pseudo-rotation angle Δθ4,5 is less than the threshold value θT, the moving image data generating unit 58 does not select the captured image [5] as the frame image [3].

[0118] If the captured image [5] is not selected as the frame image [3], the moving image data generating unit 58 determines whether or not to select the following captured image [6] as the frame image [3] .

[0119] In this case, the moving image data generating unit 58 determines whether the cumulative pseudo-rotation angle between the captured image [4] selected as the preceding frame image [2] and the captured image [6] is equal to or more than the threshold value θT. That is, the moving image data generating unit 58 determines whether the sum of the pseudo-rotation angle Δθ4,5 between the captured images [4] and [5] and the pseudo-rotation angle Δθ5,6 between the captured images [5] and [6] is equal to or more than the threshold value θT.

[0120] As shown in FIG. 13, if the sum of the pseudo-rotation angles Δθ4,5 and Δθ5,6 is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [6] as the frame image [3].

[0121] As described above, if a captured image [i] is selected as a frame image [j], the moving image data generating unit 58 determines whether the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+1] is equal to or more than the threshold value θT in order to determine whether or not to select the captured image [i+1] as the frame image [j+1].

[0122] If the cumulative pseudo-rotation angle is equal to or more than the threshold value θT, the moving image data generating unit 58 selects the captured image [i+1] as the frame image [j+1]. On the other hand, if the cumulative pseudo-rotation angle is less than the threshold value θT, the moving image data generating unit 58 does not select the captured image [i+1] as the frame image [j+1] . In this case, the moving image data generating unit 58 determines whether the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+2] is equal to or more than the threshold value θT in order to determine whether or not to select the captured image [i+2] as the frame image [j+1].

[0123] In this way, the moving image data generating unit 58 searches for a first captured image 40 having a value of the cumulative pseudo-rotation angle between itself and the captured image [i] equal to or more than the threshold value θT. For example, if the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+k] (k: an integer of 1 or more) is the threshold value θT, the moving image data generating unit 58 selects the captured image [i+k] as the frame image [j+1].

[0124] As described above, the moving image data generating unit 58 selects each frame image 42 of the moving image data.

[0125] A method for selecting the frame images 42 is not limited to the example illustrated in FIG. 13. Other examples of selecting the frame images 42 will be discussed below.

[0126] Now, the first example will be discussed. FIG. 14 is a diagram for explaining the first example. Here, a case is assumed in which, after the captured image [i] is selected as the frame image [j], if the captured image [i+1] is not selected as the frame image [j+1], it is determined whether or not to select the captured image [i+k] (k: an integer of 2 or more) as the frame image [j+1].

[0127] In this case, in the selecting method shown in FIG. 13, it is configured to determine whether the cumulative pseudo-rotation angle between the captured image [i] selected as the preceding frame image [j] and the captured image [i+k] is equal to or more than the threshold value θT. That is, in the selecting method shown in FIG. 13, it is configured to determine whether the cumulative pseudo-rotation angle Δθi,i+1+...+Δθi-k-1,i+k between the captured image [i] and the captured image [i+k] is equal to or more than the threshold value θT.

[0128] On the other hand, in the selecting method shown in FIG. 14, change amounts of the X-axis coordinate values of the feature points 60 corresponding to the certain point on the item 30 are calculated between the captured image [i] and the captured image [i+k], and based on the calculated change amounts, the pseudo-rotation angle Δθi,i+k between the captured image [i] and the captured image [i+k] is calculated.

[0129] That is, for example, in a case where a plurality of feature points 60 respectively corresponding to certain points on the item 30 are specified in the observation areas 110 in both of the captured image [i] and the captured image [i+k], an average value of the change amounts of the X-axis coordinate values of the feature points 60 is calculated as the pseudo-rotation angle Δθi,i+k, and it is determined whether the pseudo-rotation angle Δθi,i+k is equal to or more than the threshold value θT.

[0130] The second example will be discussed. Here, a case is also discussed in which, after the captured image [i] is selected as the frame image [j], it is determined whether or not to select the captured image [i+k] as the frame image [j+1].

[0131] In this case, it may be determined whether the cumulative pseudo-rotation angle between the captured image [1] selected as the first frame image [1] and the captured image [i+k] (or pseudo-rotation angle θl,i+k) is equal to or more than θT*j. Then, if the cumulative pseudo-rotation angle (or the pseudo-rotation angle θι,i+k) is equal to or more than θT*j, the captured image [i+k] may be selected as the frame image [j+1].

[0132] The third example will be discussed. FIG. 15 is a diagram for explaining the third example. Here, a case is assumed in which, after the captured image [i] is selected as the frame image [j], it is determined whether or not to select the captured image [i+3] as the frame image [j+1].

[0133] In the selecting method shown in FIG. 13, in a case where the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+3] is equal to or more than the threshold value θT, the captured image [i+3] is selected as the frame image [j+1]. Further, in the selecting method shown in FIG. 14, in a case where the pseudo-rotation angle θi,i+3 between the captured image [i] and the captured image [i+3] is equal to or more than the threshold value θT, the captured image [i+3] is selected as the frame image [j+1]. Also in the second example, in a case where the cumulative pseudo-rotation angle between the captured image [1] and the captured image [i+3] (or pseudo-rotation angle θl,i+3) is equal to or more than θT*j, the captured image [i+3] is selected as the frame image [j+1].

[0134] However, in this case, either one of the captured image [i+3] and the captured image [i+2] immediately before the captured image [i+3] may be selectively selected as the frame image [j+1].

[0135] For example, in the case of the selecting method shown in FIG. 13, in the above described case, the moving image data generating unit 58 determines which one of the cumulative pseudo-rotation angle Δθi,i+1+Δθi+1,i+2 between the captured image [i] and the captured image [i+2] and the cumulative pseudo-rotation angle Δθi,i+1+Δθi+1,i+2+Δθi+2,i+3 between the captured image [i] and the captured image [i+3] is closer to the threshold value θT. The moving image data generating unit 58 determines, for example, whether the intermediate value of these cumulative pseudo-rotation angles is less than the threshold value θT. In other words, the moving image data generating unit 58 determines whether the intermediate value of the cumulative pseudo-rotation angles, ((Δθi,i+1+Δθi+1,i+2)+(Δθi,i+1+Δθi+1,i+2+Δθi+2,i+3))/2, is less than the threshold value θT.

[0136] As shown in FIG. 15, if the intermediate value is greater than the threshold value θT, the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+2] is closer to the threshold value θT, and thus the moving image data generating unit 58 selects the captured image [i+2] as the frame image [j+1]. On the other hand, if the intermediate value is less than the threshold value θT, the cumulative pseudo-rotation angle between the captured image [i] and the captured image [i+3] is closer to the threshold value θT, and thus the moving image data generating unit 58 selects the captured image [i+3] as the frame image [j+1].

[0137] Further, in the selecting method shown in FIG. 14, the moving image data generating unit 58 determines which one of the pseudo-rotation angle between the captured image [i] and the captured image [i+2], θi,i+2, and the pseudo-rotation angle between the captured image [i] and the captured image [i+3], θi,i+3, is closer to the threshold value θT. For example, the moving image data generating unit 58 determines whether the intermediate value of these pseudo-rotation angles, (Δθi,i+2+Δθi,i+3)/2, is less than the threshold value θT.

[0138] As shown in FIG. 15, if the intermediate value is greater than the threshold value θT, the pseudo-rotation angle between the captured image [i] and the captured image [i+2], θi,i+2, is closer to the threshold value θT, and thus the moving image data generating unit 58 selects the captured image [i+2] as the frame image [j+1] . On the other hand, if the intermediate value is less than the threshold value θT, the pseudo-rotation angle between the captured image [i] and the captured image [i+3], θi,i+3, is closer to the threshold value θT, and thus the moving image data generating unit 58 selects the captured image [i+3] as the frame image [j+1].

[0139] In this way, the item 30 rotates more smoothly in the moving image data. If the intermediate value is equal to the threshold value θT, the moving image data generating unit 58 may select either one of the captured image [i+2] and the captured image [i+3] as the frame image [j+1].

[0140] Next, the processing executed in the image processing device 10 to implement the functional blocks shown in FIG. 5 will be discussed. FIGs. 16 and 17 are flow charts showing an example of the processing executed in the image processing device 10 in order to generate moving image data of the item 30 rotating once based on captured data. The processing shown in FIGs. 16 and 17 corresponds to a selecting methods made by combining the selection methods shown in FIGs. 13 and 15. The control unit 11 executes the processing shown in FIGs. 16 and 17 according to a program, to thereby operate as the functional blocks shown in FIG. 5.

[0141] As shown in FIG. 16, the control unit 11 extracts captured image data corresponding to one rotation of the item 30 from the captured image data in order to generate moving image data of one rotation of the item 30 (S101) . FIG. 18 is a diagram illustrating an outline of the processing of step S101, and FIG. 19 is a flow chart showing details of the processing of step S101.

[0142] As shown in FIG. 18, the control unit 11 searches for a captured image 40 at the point of time when the item 30 completes one rotation from among the captured images 40 included in the captured image data. The captured image 40 at the point of time when the item 30 completes one rotation is supposed to be an image that is the same or substantially the same as the captured image 40 immediately before the item 30 starts to rotate (hereinafter "reference captured image") and is shot after a certain period of time has passed since the reference captured image is shot. As such, if the captured image [i] satisfies both of the following conditions (A) and (B), it can be assumed that the captured image [i] corresponds to the captured image 40 at the point of time when the item 30 completes one rotation. For example, the first captured image 40 from among the captured images 40 included in the captured image data is set as "reference captured image."
  1. (A) The reference captured image is the same or substantially the same as the captured image [i]. In other words, the similarity between the reference captured image and the captured image [i] is extremely high.
  2. (B) The point of time when the reference captured image is shot and the point of time when the captured image [i] is shot are separated from each other for a predetermined period of time or more. In other words, a difference between the image number of the reference captured image and the image number of the captured image [i] is equal to or more than the threshold value (IT).


[0143] "Predetermined period of time" and "threshold value (IT)" in the above condition (B) are determined in view of an average time required for the item 30 to complete one rotation.

[0144] As described above, the control unit 11 searches for a captured image 40 that satisfies both of the conditions (A) and (B) from among the captured image data, to thereby find out a captured image 40 at the point of time when the item 30 completes one rotation. The control unit 11 then extracts images, starting from the reference captured image to such a captured image 40, as captured image data of one rotation of the item 30. In the processing shown in FIG. 19, a captured image 40 that satisfies both of the conditions (A) and (B) is searched in the order from the last captured image 40 to the first captured image 40. Here, the first captured image 40 is set as the reference captured image.

[0145] In the processing shown in FIG. 19, the control unit 11 firstly obtains the total number N of captured images 40 included in the captured image data (S201). The control unit 11 then substitutes the total number N for a variable i (S202). Further, the control unit 11 set a value obtained by multiplying the total number N by a predetermined coefficient α (α: 0<α<1, e.g., 1/3) as a threshold value IT (S203) . The control unit 11 may set a predetermined fixed value as a threshold value IT.

[0146] Subsequently, the control unit 11 determines whether the captured image [1] and the captured image [i] are the same or substantially the same with each other (S204) . That is, the control unit 11 determines whether the similarity between the captured image [1] and the captured image [i] is extremely high.

[0147] For example, in a case where a plurality of feature points 60 respectively corresponding to certain points on the item 30 are specified in both of the captured image [1] and the captured image [i], the control unit 11 determines whether the number of such feature points 60 is greater than a predetermined threshold value.

[0148] If the number is greater than the threshold value, the control unit 11 generates a subtraction image between the captured image [1] and the captured image [i], and calculates a total value of pixel values of the subtraction image. In this case, when the total value is smaller, the similarity between the captured image [1] and the captured image [i] is higher. As such, the control unit 11 determines whether the total value is less than the predetermined threshold value.

[0149] If the total value is less than the threshold value, the control unit 11 determines that captured image [1] and the captured image [i] are the same or substantially the same. On the other hand, when the number is less than the threshold value or when the total value is equal to or more than the threshold value, the control unit 11 determines that the captured image [1] and the captured image [i] are not the same or substantially the same.

[0150] In the determining method as described above, for example, a subtraction image is generated only when the above number is greater than the threshold value, and, as a result, the processing load can be reduced. However, the method for determining whether the captured image [1] and the captured image [i] are the same or substantially the same is not limited to the above method. In the above described method, it is configured to consider the number of feature points 60, but it may be configured not to consider the number of feature points 60. Alternatively, a known method other than the above described method may also be applied.

[0151] If the captured image [1] and the captured image [i] are not the same or substantially the same (S204: N), the control unit 11 subtracts 1 from the value of variable i (S205), and determines whether the value of variable i is less than the threshold value IT (S206) . If the value of variable i is not less than the threshold value IT (S206: N), the control unit 11 executes step S204.

[0152] On the other hand, if the captured image [1] and the captured image [i] are the same or substantially the same (S204: Y), the control unit 11 extracts captured images [1] to [i] as captured image data of one rotation (S208). That is, the control unit 11 uses the captured images [1] to [i] in the processing described later (S102-S120) . In this case, captured images 40 from and after the captured image [i+1] are removed.

[0153] In step S206, in a case where it is determined that the value of variable i is less than the threshold value IT, the control unit 11 extracts the captured images [1] to [N] (S207) . The case where it is determined that the value of variable i is less than the threshold value IT in step S206 is a case where there is no captured image 40 that satisfies both of the conditions (A) and (B) . In this case, the control unit 11 uses all of the captured images 40 included in the captured image data in the processing described later (S102-S120).

[0154] Instead of executing steps S204-S208, a captured image 40, the similarity between which and the captured image [1] is the highest among the captured images [IT] to [N], may be searched. Then, the captured images, starting from the first captured image 40 to the captured image 40 that has been found, may be extracted as data of one rotation.

[0155] As shown in FIG. 16, after step S101 is executed, the control unit 11 reads the captured image [1] (S102), and specifies feature points 60 in the captured image [1] (S103). Further, the control unit 11 selects the captured image [1] as the frame image [1] (S104) .

[0156] Subsequently, the control unit 11 initializes the variable s to 0, the variables i and j to 2, and the variable k to 1 (S105) . The variable s is used to calculate a cumulative pseudo-rotation angle. The variable i represents an image number of a captured image 40 to be processed, and the variable j represents a frame number of a frame image 42 to be selected. The variable k represents an image number of the captured image 40 selected as the frame image [j-1] .

[0157] After step S105 is executed, the control unit 11 reads the captured image [i] (S106), and specifies feature points 60 in the captured image [i] (S107). A result of the specification in step S107 is stored in the storage unit 12.

[0158] After step S107 is executed, the control unit 11 calculates the pseudo-rotation angle Δθi-1,i between the captured image [i-1] and the captured image [i] (S108), and adds the pseudo-rotation angle Δθi-1,i to the variable s (S109). The control unit 11 then determines whether the variable s is equal to or more than the threshold value θT (S110).

[0159] If the variable s is not equal to or more than the threshold value θT (S110: N), as shown in FIG. 17, the control unit 11 adds 1 to the variable i (S119), and determines whether there is a captured image [i] (S120) . If there is a captured image [i] (S120: Y), the control unit 11 executes step S106.

[0160] On the other hand, if the variable s is equal to or more than the threshold value θT (S110: Y), as shown in FIG. 17, the control unit 11 calculates an intermediate value, (sk,i-1+sk,i)/2, of the cumulative pseudo-rotation angle sk,i-1 between the captured image [k] selected as the frame image [j] and the captured image [i-1] and the cumulative pseudo-rotation angle sk,i between the captured image [k] and the captured image [i], and determines whether the intermediate value is equal to or less than the threshold value θT (Sill).

[0161] If the intermediate value is equal to or less than the threshold value θT (S111: Y), the control unit 11 selects the captured image [i] as the frame image [j] (S112). In this case, the control unit 11 substitute the value of variable i into the variable k (S113), and adds 1 to the variable i (S114) . The control unit 11 then determines whether there is a captured image [i] (S115) . If there is a captured image [i] (S115: Y), the control unit 11 adds 1 to the variable j (S118), and executes step S106.

[0162] On the other hand, if the intermediate value is greater than the threshold value θT (S111: N), the control unit 11 selects the captured image [i-1] as the frame image [j] (S116) . In this case, the control unit 11 substitutes the value, which is obtained by subtracting 1 from the value of variable i, into the variable k (S117). The control unit 11 then adds 1 to the variable j (S118), and executes step S106.

[0163] If it is determined that there is no captured image [i] in step S115 or S120, the control unit 11 terminates the processing. In this case, moving image data including the frame images [1] to [j] is generated as moving image data that shows the item 30 that rotates once.

[0164] The processing executed in the image processing device 10 is not limited to the processing shown in FIGs. 16 and 17.

[0165] For example, steps S111, S116, and S117 in FIG. 17 may be removed. That is, it may be configured that if it is determined that the variable s is equal to or more than the threshold value θT in step S110, steps S112 to S115 are always executed. In this case, the selecting method shown in FIG. 13 is executed.

[0166] Moreover, for example, instead of executing steps S108 and S109 in FIG. 16, the control unit 11 may calculate a pseudo-rotation angle between the captured image [k] and the captured image [i], θk,i. And in step S110, it may be determined whether the pseudo-rotation angle θk,i is equal to or more than the threshold value θT. In this case, the selecting method shown in FIG. 14 is executed.

[0167] Alternatively, for example, step S101 in FIG. 16 may be skipped and the processing shown in FIG. 20 may be executed instead of the processing shown in FIG. 17. The processing shown in FIG. 20 is different from the processing shown in FIG. 17 at the point that it includes steps S121 to S124.

[0168] In the processing shown in FIG. 20, after steps S113 and S117 are executed, the control unit 11 executes processing for determining whether the captured image 40 selected as the frame image [j] corresponds to the captured image 40 at the point of time when the item 30 has completed one rotation (steps S121-S124).

[0169] That is, after step S113 is executed, the control unit 11 determines whether the value of variable i is equal to or more than the threshold value IT (S121) . The threshold value IT is the same as the one in step S206 of FIG. 19.

[0170] If the value of variable i is not equal to or more than the threshold value IT (S121: N), the control unit 11 executes step S114. On the other hand, if the value of variable i is equal to or more than the threshold value IT (S121: Y), the control unit 11 determines whether the frame image [1] and the frame image [j] are the same or substantially the same(S122). This processing is the same as the one in step S204 of FIG. 19.

[0171] If it is determined that the frame image [1] and the frame image [j] are not the same or substantially the same (S122: N), the control unit 11 executes step S114. On the other hand, if it is determined that the frame image [1] and the frame image [j] are the same or substantially the same(S122: Y), the control unit 11 terminates the processing. In this case, moving image data including the frame images [1] to [j] is generated as moving image data that shows the item 30 that rotates once.

[0172] After step S117 is executed, the control unit 11 determines whether the value obtained by subtracting 1 from the value of variable i is equal to or more than the threshold value IT (S123). If the value obtained by subtracting 1 from the value of variable i is not equal to or more than the threshold value IT (S123: N), the control unit 11 executes step S118. On the other hand, if the value obtained by subtracting 1 from the value of variable i is equal to or more than the threshold value IT (S123: Y), the control unit 11 determines whether the frame image [1] and the frame image [j] are the same or substantially the same (S124).

[0173] If it is determined that the frame image [1] and the frame image [j] are not the same or substantially the same (S124: N), the control unit 11 executes step S118. On the other hand, if it is determined that the frame image [1] and the frame image [j] are the same or substantially the same (S124: Y), the control unit 11 terminates the processing. In this case, moving image data including the frame images [1] to [j] is generated as moving image data that shows the item 30 that rotates once.

[0174] With the manner described above, moving image data of the item 30 that rotates once can also be generated.

[0175] Instead of executing steps S121 to S124, it may be configured that the processing shown in FIG. 21 is executed in a case where it is determined that there is no captured image [i] in step S115 or S120. The processing shown in FIG. 21 is processing for extracting moving image data of one rotation from among the frame images [1] to [j] included in the generated moving image data.

[0176] In the processing shown in FIG. 21, the control unit 11 substitutes the value of variable j into the variable h (S301), and sets a value obtained by multiplying the variable j by a coefficient α (0<α<1: e.g., 1/3) as a threshold value HT (S302). The threshold value HT corresponds to the threshold value IT in FIGs. 19 and 20.

[0177] Subsequently, the control unit 11 determines whether the frame image [1] and the frame image [h] are the same or substantially the same (S303). This processing is the same as the one in step S204 of FIG. 19.

[0178] If the frame image [1] and the frame image [h] are not the same or substantially the same (S303: N), the control unit 11 subtracts 1 from the value of variable h (S304) . The control unit 11 then determines whether the value of variable h is less than the threshold value HT (S305). If the value of variable h is not less than the threshold value HT (S305: N), the control unit 11 executes step S303.

[0179] On the other hand, if the frame image [1] and the frame image [h] are the same or substantially the same (S303: Y), the control unit 11 extracts the frame images [1] to [h] as moving image data of the item 30 that rotates once (S307) . In this regard, the frame images [1] to [h] are obtained from the once selected frame images [1] to [j] as the moving image data of the item 30 that rotates once. In this case, the frame images [h+1] to [j] are removed.

[0180] In a case where the value of variable h is less than the threshold value HT (S305: Y), the control unit 11 extracts the frame images [1] to [j] (S306) . The case where it is determined that the value of variable h is less than the threshold value HT in step S306 is a case where a frame image 42 corresponding to the image of the point of time when the item 30 completes one rotation could not be found. In this case, the control unit 11 extracts the frame images [1] to [j] as moving image data of the item 30.

[0181] The moving image data of the item 30 that rotates once can be generated also by executing the processing shown in FIG. 21 instead of the processing shown in FIG. 20.

[0182] In the processing shown in FIG. 21, a frame image 42 having the highest similarly to the frame image [1] among the frame images [HT] to [j] may be searched. Then, the moving image data of the images, starting from the first frame image 42 to the frame image 42 that has been found, may be obtained as moving image data of the item 30 that rotates once. In this manner, the moving image data of the item 30 that rotates once can also be generated.

[0183] Using the image processing device 10 described above, it is possible to obtain moving image data in which a rotation angle of the item 30 is approximately constant for each frame based on captured image data obtained by shooting the item 30 rotating at a non-constant angular velocity. Further, using the image processing device 10, it is possible to obtain moving image data of one rotation.

[0184] The present invention is not to be limited to the above described embodiment.

[1] In the above, the explanation has been given based on the assumption that the shape of the item 30 is substantially cylindrical, and a radius r of an orbit 80 of each point on the item 30 (circular orbit) is all the same. However, the shape of the item 30 is not limited to a substantially cylindrical shape. In a case where the shape of the item 30 is not substantially cylindrical, the radius r varies at each point, and thus an error may possibly be large if the radius r is assumed to be all the same. As such, the image processing device 10 may have a configuration to suppress occurrence of such an inconvenience.

[1-1] For example, the rotation angle information obtaining unit 56 may be configured to use a feature point 60 having a large radius r in preference to a feature point 60 having a small radius r in order to obtain rotation angle information.

[1-1-1] For example, the rotation angle information obtaining unit 56 may be configured to use only the feature point 60 having a large radius r to obtain rotation angle information, and not to use the feature point 60 having a large radius r, to obtain rotation angle information.
For example, it can be said that a feature point 60 that passes through both inside and outside the observation area 110 corresponds to a point on the item 30 (feature point 60) having a relatively large radius r. As such, it can be configured that, by using only such a feature point 60 to obtain rotation angle information, only a feature point 60 having a relatively large radius r is used to obtain rotation angle information.
In this case, it is enough to determine whether the feature point 60 moves into the observation area 110 from outside the observation area 110, and not necessary to calculate the radius r, and hence an increase of the processing load can be suppressed.

[1-1-2] As a matter of course, the rotation angle information obtaining unit 56 (rotation radius information obtaining means) may be configured to calculate a radius r (rotation radius information) for each feature point 60. And, the rotation angle information obtaining unit 56 may be configured to use only a feature point 60 having a radius r larger than a threshold value in order to obtain rotation angle information, and not to use a feature point 60 having a radius r equal to or less than a threshold value in order to obtain rotation angle information.

[1-2] Further, for example, the rotation angle information obtaining unit 56 (rotation radius information obtaining means) may calculate a radius r (rotation radius information) for each feature point 60. And, the rotation angle information obtaining unit 56 may calculate a pseudo-rotation angle (rotation angle information) of a feature point 60 in accordance with the equation (11), based on a change amount of the X-axis coordinate value of the feature point 60 (position difference information) and the radius r of the feature point 60 in question (rotation radius information) . Then, the rotation angle information obtaining unit 56 may use a statistical value (e.g., average value) of the pseudo-rotation angle calculated this way as the pseudo-rotation angle of the item 30.


In the variations 1-1-2 and 1-2 described above, a radius r needs to be calculated. As such, a method for calculating the radius r (rotation radius information) will be discussed.
For example, the rotation angle information obtaining unit 56 (rotation radius information obtaining means) may obtain elliptical orbit information relating to an orbit 70 (elliptical orbit) of a feature point 60 corresponding to a certain point on the item 30, and calculate the radius r based on the elliptical orbit information. For example, the rotation angle information obtaining unit 56 estimates the radius r based on at least one of major and minor diameters of the orbit 70. The rotation angle information obtaining unit 56 obtains, for example, one half of the major diameter of the orbit 70 as the radius r.
If the feature point 60 corresponding to the certain point on the item 30 is specified in at least three of the captured images 40, elliptical orbit information relating to the orbit 70 of the feature points 60 can be obtained. In this case, an accuracy of calculating the orbit 70 may be enhanced as described below.
In a case where an orbit 70 is calculated for each of the feature points 60 respectively corresponding to the certain points on the item 30, the center points of these orbits 70 match and they should be concentric ellipses. As such, the rotation angle information obtaining unit 56 can be configured so as not to use an orbit 70 that has a center point not matching those of the other orbits 70 and is not concentric with other orbits 70.
If the feature point 60 corresponding to a certain point on the item 30 is specified in four or more of the captured images 40, the rotation angle information obtaining unit 56 may obtain a plurality of combinations of three positions of the feature points 60. Further, the rotation angle information obtaining unit 56 may calculate a plurality of orbits 70 respectively corresponding to the combinations. The rotation angle information obtaining unit 56 then may obtain an average of each of the center points and major and minor diameters of the orbits 70 calculated as above as the center point and major and minor diameters of the orbit 70 of the feature point 60. In this way, it is also possible to enhance an accuracy of calculating the orbit 70.
The following method may also be adapted instead of the method described above.
For example, in a case where the feature point 60 corresponding to a certain point on the item 30 is first specified in a captured image [i] (i: an integer of 1 or more) and specified last in a captured image [j] (j: an integer greater than i), the rotation angle information obtaining unit 56 (rotation radius information obtaining means) may obtain the radius r based on positions of the feature point 60 in the captured images [i] and [j].
For example, if the positions of the feature point 60 in the captured images [i] and [j] are both included in an edge area of the existence area 100, the rotation angle information obtaining unit 56 may obtain one half of the length of a straight line connecting the positions of the feature point 60 in question in the captured images [i] and [j] as the radius r. In this regard, the "edge area" is an area within a predetermined distance from the edge of the existence area 100.
Alternatively, the rotation angle information obtaining unit 56 may obtain one half of the difference in the X-axis coordinate values of the feature point 60 between the captured images [i] and [j] as the radius r.
By the way, the pattern on the item 30 deforms around the edge of the existence area 100, and thus a feature amount of a feature point 60 corresponding to the pattern part is different between a case where the feature point 60 is positioned at the center area of the existence area 100 and a case where the feature point 60 is positioned at the edge area of the existence area 100. Accordingly, in a case where the feature point 60 corresponding to the pattern part is positioned at the edge area of the existence area 100, such a feature point 60 could be recognized as being different from the feature point 60 corresponding to the pattern part. Further, when the shooting device 34 and the item 30 are close, there is a case where 180 degrees rotation of the pattern part of the item 30 is not imaged, and only a rotation of an angle less than 180 degrees is imaged. Considering this, the calculation method mentioned previously has an accuracy higher than that of this calculation method. However, this calculation method does not require a calculation of an elliptical orbit, and thus can reduce the processing load compared to the calculation method mentioned previously.

[2] In the above described embodiment, for example, the threshold value θT is set in advance, but the threshold value θT may be dynamically set.

[2-1] For example, the threshold value θT may be determined, based on the number of captured images 40 that serve as a basis for generating moving image data. The threshold value θT may be set in a way that, as the number of the captured images 40 corresponding to one rotation of the item 30 is greater, the threshold value θT is smaller. In this case, it may be configured that correspondence relation information indicating the correspondence relation of the number of captured images 40 and the threshold value θT is stored in the storage unit 12, and this threshold value θT is set based on the correspondence relation information.
As the threshold value θT is smaller, the moving image data becomes smoother. In this regard, with the manner described above, the threshold value θT can be reduced in accordance with captured data, and thus the moving image data can be smoothed as much as possible.

[2-2] Further, for example, the threshold value θT may be changed, when frame images 42 are not selected enough. FIG. 22 is a diagram for explaining this variation.
In this variation, firstly, the control unit 11 sets the threshold value θT to a relatively small value. The control unit 11 then selects frame images 42 so that a rotation angle between the adjacent frame images 42 is the threshold value θT.
In this case, the control unit 11 determines whether the frame images 42 have been selected so that the rotation angle between the adjacent frame images 42 is the threshold value θT. For example, if a difference between the pseudo-rotation angle between the frame image [j] and the frame image [j+1] and the threshold value θT is within a predetermined range, the control unit 11 assumes that the frame images 42 have been selected so that the rotation angle between the adjacent frame images 42 is the threshold value θT. Specifically, if the pseudo-rotation angle is within a range of plus or minus 20% of the threshold value θT, the control unit 11 assumes that the frame images 42 have been selected so that the rotation angle between the adjacent frame images 42 is the threshold value θT.
Further, in this case, as shown in the frame images [2] and 6 in FIG. 22 (A), when the frame images 42 have not been selected so that the rotation angle between the adjacent frame images 42 is the threshold value θT, the control unit 11 increases the threshold value θT as shown in FIG. 22(B), and selects the frame images 42 based on the updated threshold value θT. Such a processing is repeated until all of the frame images 42 are selected so that the rotation angle between the adjacent frame images 42 is the threshold value θT.
In this way, the threshold value θT can also be reduced in accordance with captured image data, and thus the moving image data can be smoothed as much as possible.

[2-3] It may be configured that the captured image data corresponding to multiple rotations is obtained in advance, and a frame image 42, which has not been obtained from the captured image data of the first rotation, is obtained from the captured image data of the second and following rotations. The captured data of the second and following rotations can be extracted in the same way as the captured data of the first rotation is extracted.
For example, in the example shown in FIG. 22(A), the frame image [2], the pseudo-rotation angle between which and the frame image [1] is ΔθT, has not been selected. In this case, a frame image 42, the pseudo-rotation angle between which and the frame image [1] is ΔθT, may be searched from the captured image data of the second and following rotations.
Then, it may be configured that the threshold value θT is increased as shown in FIG. 22(B) only when a frame image 42, the pseudo-rotation angle between which and the frame image [1] is ΔθT, cannot be found from the captured image data of the second and following rotations.

[3] For example, in a case where the shape of item 30 is rotationally symmetric to the rotation axis 32, and only a part of the item 30 (e.g., obverse side) has letters, designs, or patterns, the captured images 40 does not change while the part with letters etc. is not covered by the shooting device 34. In this case, a feature point 60 corresponding to a certain point on the item 30 cannot be specified, and as a result, a change amount of X-axis coordinate values of the feature point 60 cannot be obtained.
FIG. 23 is a diagram for explaining the processing for addressing the inconvenience as mentioned above. FIG. 23 illustrates a case where a feature point 60 corresponding to a certain point on the item 30 is not specified in captured images [m+1] to [m+n] from among captured images [1] to [m+n] included in the captured image data of one rotation.
In this case, the control unit 11 obtains the number of captured images 40 that are selected from the captured images [1] to [m] as frame images 42. In the example of FIG. 23, the number is p.
Then the control unit 11 selects the frame images 42 from the captured images [m+1] to [m+n] based on a ratio of the number of the captured images [1] to [m] to the number of the captured images 40 that are selected from the captured images [1] to [m] as frame images 42.
In the example shown in FIG. 23, the ratio is p/m, and thus the control unit 11 selects the frame images 42 from the captured images [m+1] to [m+n] at the same ratio as this ratio. In other words, the control unit 11 selects q (q: q=n*p/m) captured images 40 from n captured images [m+1] to [m+n] as frame images 42. For example, the control unit 11 selects a captured image 40 as a frame image 42 at intervals of m/p from among the captured images [m+1] to [m+n] . If m/p is "3", the control unit 11 selects captured images [m+3], [m+6], and [m+9],... as frame images 42. In this way, it is possible to address the inconvenience as described above.
In the above, a description has been given regarding a case where a feature point 60 corresponding to a certain point on the item 30 is not specified in the captured images [m+1] to [m+n], and the same description can be applied to a case where a feature point 60 corresponding to a certain point on the item 30 is not specified in the captured images [1] to [m]. In this case, the control unit 11 may select frame images 42 from the captured images [1] to [m] based on a ratio of the number of the captured images [m+1] to [m+n] to the number of the captured images 40 that are selected from the captured images [m+1] to [m+n] as frame images 42.

[4] Further, for example, operations of the rotation angle information obtaining unit 56 are not limited to the embodiment described above.



[0185] For example, the rotation angle information obtaining unit 56 may obtain elliptical orbit information relating to the orbit 70 of the feature point 60 (elliptical orbit) corresponding to the certain point on the item 30. If a feature point 60 corresponding to a certain point on the item 30 is specified in at least three captured images 40, it is possible to obtain elliptical orbit information of the orbit 70 of the feature point 60 based on the positions of the feature point 60 on these captured images 40.

[0186] In a case where, for example, a rotation angle of the item 30 between the point of time when the captured image [i] is shot and the point of time when the captured image [i+k] (k: an integer of 1 or more) is shot is calculated, the rotation angle information obtaining unit 56 may calculate the rotation angle based on the position of the feature point 60 on the captured image [i], the position of the feature point 60 on the captured image [i+k], and the elliptical orbit information. In this manner, the rotation angle information of the rotation angle of the item 30 between the point of time when the captured image [i] is shot and the point of time when the captured image [i+k] is shot can also be obtained.


Claims

1. An image processing device (10) comprising:

captured image obtaining means (50) for reading captured image data of a plurality of captured images that are obtained by repeatedly shooting an object (30) while the object (30) completes one rotation;

feature point specifying means (52) for specifying, using an algorithm, a plurality of feature points (60) in the plurality of captured images (40) which were shot while the object completed the one rotation,

position information obtaining means (54) for obtaining, in a case where a feature point (60) corresponding to a certain point on the object (30) is specified in a first captured image and a second captured image from among the plurality of captured images, the second captured image being shot after the first captured image and shot before a point of time when the object completes the one rotation, position information (x, y) relating to positions of the feature point (60) in the first captured image and the second captured image;

rotation angle information obtaining means (56) configured to calculate a radius (r) for each feature point (60) based on the captured images, for obtaining rotation angle information (θ1, θ2) relating to a rotation angle of the object between a point of time when the first captured image is shot and a point of time when the second captured image is shot, based on the position information and the calculated radius; and

moving image data generating means (58) for generating moving image data (42) which shows how the object rotates (30), by selecting a plurality of frame images of the moving image data (42) from among the plurality of captured images (40), based on the rotation angle information (θ1, θ2) obtained by the rotation angle information obtaining means for each of the plurality of captured images (40).


 
2. The image processing device (10) according to claim 1,
wherein the rotation angle information obtaining means (56) comprises:

means for obtaining position difference information (x1, x2) relating to a difference between the position of the feature point in the first captured image and the position of the feature point in the second captured image; and

means for obtaining the rotation angle information (θ1, θ2) based on the position difference information (x1, x2).


 
3. The image processing device (10) according to claim 2 comprising:

means for setting an area in the captured image as an observation area (110) based on a position of a rotation axis (32) of the object (30) in the captured image,

wherein the rotation angle information obtaining means (56) is configured to only use the feature point (60) to obtain the rotation angle information (θ1, θ2) based on the position difference information (x1, x2) if the feature point (60) is included in the observation area (110).


 
4. The image processing device (10) according to claim 2 or 3,
wherein the rotation angle information obtaining means (56) comprises:

rotation radius information obtaining means for obtaining, in a case where a plurality of feature points (60) respectively corresponding to a plurality of points on the object are specified in the first captured image (40) and the second captured image (40), rotation radius information relating to a rotation radius of a point on the object (30) corresponding to the feature point (60), for each of the plurality of feature points (60);

means for obtaining, for each of the plurality of feature points (60), rotation angle information relating to a rotation angle of the point corresponding to the feature point (60) on the object (30) between a point of time when the first captured image is shot and a point of time when the second captured image is shot, based on the position difference information (x1, x2) and the rotation radius information of the feature points; and

means for obtaining the rotation angle information of the object based on the rotation angle information of each of the plurality of feature points.


 
5. The image processing device (10) according to claim 4,
wherein the rotation angle information obtaining means (56) obtains the rotation angle information (θ1, θ2) using the position difference information (x1, x2) of a feature point (60) with a first rotation radius when the object (30) rotates, in preference to the position difference information of a feature point (60) with a second rotation radius when the object (30) rotates which is smaller than the first rotation radius.
 
6. The image processing device (10) according to claim 4,
wherein the rotation radius information obtaining means comprises:

means for obtaining, in a case where the feature point (60) corresponding to the certain point on the object (30) is specified in at least three captured images (40), elliptical orbit information relating to an elliptical orbit of the feature point (60) based on the position of the feature point (60) in the at least three captured images (40); and

means for obtaining the rotation radius information of the feature point (60) based on the elliptical orbit information of the feature point (60).


 
7. The image processing device (10) according to claim 1,
wherein the rotation angle information obtaining means (56) comprises:

means for obtaining, in a case where the feature point (60) corresponding to the certain point on the object (30) is specified in at least three captured images, elliptical orbit information relating to an elliptical orbit of the feature point (60); and

means for obtaining the rotation angle information based on the positions of the feature point (60) in the first captured image and in the second captured image and the elliptical orbit information.


 
8. The image processing device (10) according to any one of claims 1 to 7,
wherein the moving image data generating means (58) comprises:

means for obtaining similarity information relating to a similarity between a reference captured image (40) that is selected from the plurality of captured images (40) and each of a plurality of captured images (40) that are shot after a predetermined period of time passes from a point of time when the reference captured image (40) is shot; and

means for selecting one of the plurality of captured images (40) that are shot after the predetermined period of time passes from the point of time when the reference captured image (40) is shot as a captured image (40) of a point of time when the object (30) completes one rotation based on the similarity information, and

wherein the moving image data generating means (58) generates the moving image data (42) showing how the object rotates once based on captured images (40) that are shot before the point of time when the captured image (40) selected as the captured image (40) of the time when the object (30) completes one rotation is shot.
 
9. The image processing device (10) according to claim 8,
wherein the moving image data generating means (58) comprises:

feature point number obtaining means for obtaining a number of feature points (60) corresponding to feature points (60) specified in the reference captured image (40), from among the feature points (60) specified in the captured image (40) that is shot after the predetermined period of time passes from the point of time when the reference captured image (40) is shot;

means for selecting candidates of the captured image (40) of the point of time when the object (30) completes one rotation, from among the plurality of captured images (40) that are shot after the predetermined period of time passes from the point of time when the reference captured image (40) is shot, based on a result obtained by the feature point number obtaining means;

means for obtaining similarity information relating to a similarity between the reference captured image (40) and each of the captured images (40) selected as the candidates; and

means for selecting one of the captured images (40) selected as the candidates as the captured image (40) of a point of time when the object (30) completes one rotation based on the similarity information.


 
10. The image processing device (10) according to any one of claims 1 to 9,
wherein the plurality of captured images (40) include the 1st to mth (m: an integer of 2 or more) captured images and m+1th to m+nth (n: an integer of 1 or more) captured images, which are shot after the mth captured image,
wherein the moving image data generating means (58) comprises:

selection number obtaining means for obtaining a number of the captured images (40) selected as frame images of the moving image data (42) out of the 1st to the mth captured images; and

means for selecting the frame images of the moving image data (42) out of the m+1th to the m+nth captured images, in a case where the feature point (60) corresponding to the point on the object (30) is not specified in the m+1th to the m+nth captured images, based on a ratio of a number of the 1st to the mth captured images to the number of the captured images (40) obtained by the selection number obtaining means.


 
11. The image processing device (10) according to any one of claims 1 to 10, comprising:

first determination means for determining whether or not a feature amount of the feature point (60) specified in the first captured image and a feature amount of the feature point (60) specified in the second captured image are the same or substantially the same;

second determination means for determining whether or not a difference of a coordinate value of a direction of an axis corresponding to a rotation axis (32) of the object (30) between the position of the feature point (60) specified in the first captured image and the position of the feature point (60) specified in the second captured image is equal to or more than the threshold value; and

means for determining whether or not the feature point (60) specified in the first captured image and the feature point (60) specified in the second captured image are feature points corresponding to the certain point on the object (30) based on a determination result of the first determination means and the second determination means.


 
12. The image processing device (10) according to any one of claims 1 to 11, comprising:

determination means for determining whether or not the feature amount of the feature point (60) specified in the first captured image and the feature amount of the feature point (60) specified in the second captured image are the same or substantially the same;

comparing means for comparing (a) information relating to a difference of a coordinate value of a direction of a second axis perpendicular to a first axis corresponding to a rotation axis (32) of the object (30) between the position of a first feature point in the first captured image and the position of the first feature point in the second captured image with (b) information relating to a difference of a coordinate value of the direction of the second axis between the position of a second feature point in the first captured image and the position of the second feature point in the second captured image, in a case where the first feature point having the same or substantially the same feature amount in both of the first captured image and the second captured image, and the second feature point having the same or substantially the same feature amount in both of the first captured image and the second captured image are specified; and

means for determining whether or not the feature point specified in the first captured image and the feature point specified in the second captured image are feature points (60) corresponding to the certain point on the object (30), based on a determination result of the determination means and a comparing result of the comparing means.


 
13. The image processing device (10) according to any one of claims 1 to 12,
wherein the captured image obtaining means (50) obtains a plurality of captured images (40) that are obtained by repeatedly shooting the object (30) while the object (30) rotates a plurality of times, and
wherein the moving image data generating means (58) generates the moving image data (42) that shows how the object (30) rotates once based on the plurality of captured images (40) that are obtained by repeatedly shooting the object (30) while the object (30) rotates the plurality of times.
 
14. An image processing method comprising:

a captured image obtaining step (S102) of reading captured image data of a plurality of captured images (40) that are obtained by repeatedly shooting an object (30) while the object completes one rotation;

a feature point specifying step (S103) of specifying, using an algorithm, a plurality of feature points (60) in the plurality of captured images (40) which were shot while the object completed one rotation;

a position information obtaining step of obtaining, in a case where a feature point (60) corresponding to a certain point on the object (30) is specified in a first captured image and a second captured image from among the plurality of captured images, the second captured image being shot after the first captured image and shot before a point of time when the object completes the one rotation, position information (x, y) relating to positions of the feature point (60) in the first captured image and the second captured image;

a rotation angle information obtaining step of calculating a radius (r) for each feature point based on the captured images, and of obtaining rotation angle information (θ1, θ2) relating to a rotation angle of the object (30) between a point of time when the first captured image is shot and a point of time when the second captured image is shot, based on the position information (x, y) and the calculated radius (r); and

a moving image data generating step of generating moving image data (42) by selecting a plurality of frame images of the moving image data (42), which shows how the object (30) rotates, from among the plurality of captured images (40), based on the rotation angle information (θ1, θ2) obtained by the rotation angle information obtaining step for each of the plurality of captured images (40).


 
15. A program for causing a computer to function as:

captured image obtaining means (50) for reading captured image data of a plurality of captured images (40) that are obtained by repeatedly shooting an object (30) while the object completes one rotation;

feature point specifying means (52) for specifying, using an algorithm, a plurality of feature points (60) in the plurality of captured images which were shot while the object completed one rotation,

position information obtaining means (54) for obtaining, in a case where a feature point (60) corresponding to a certain point on the object (30) is specified in a first captured image and a second captured image from among the plurality of captured images, the second captured image being shot after the first captured image and shot before a point of time when the object completes the one rotation, position information (x, y) relating to positions of the feature point in the first captured image and the second captured image;

rotation angle information obtaining means (56) configured to calculate a radius (r) for each feature point (60) based on the captured images, for obtaining rotation angle information (θ1, θ2) relating to a rotation angle of the object (30) between a point of time when the first captured image is shot and a point of time when the second captured image is shot, based on the position information (x, y) and the calculated radius (r); and

moving image data generating means (58) for generating moving image data (42) by selecting a plurality of frame images of the moving image data (42), which shows how the object (30) rotates, from among the plurality of captured images (40), based on the rotation angle information (θ1, θ2) obtained by the rotation angle information obtaining means (56) for each of the plurality of captured images (40).


 


Ansprüche

1. Bildverarbeitungsvorrichtung (10), umfassend:

Mittel zum Erhalten aufgenommener Bilder (50), um aufgenommene Bilddaten einer Vielzahl von aufgenommenen Bildern zu lesen, die erhalten werden, indem ein Objekt (30) wiederholt fotografiert wird, während das Objekt (30) eine Rotation ausführt;

Mittel zum Vorgeben von Merkmalspunkten (52), um unter Verwendung eines Algorithmus eine Vielzahl von Merkmalspunkten (60) in der Vielzahl von aufgenommenen Bildern (40), die fotografiert wurden, während das Objekt die eine Rotation ausführte, vorzugeben,

Mittel zum Erhalten von Positionsinformationen (54), um in einem Fall, in dem ein Merkmalspunkt (60), der einem gewissen Punkt an dem Objekt (30) entspricht, in einem ersten aufgenommenen Bild und einem zweiten aufgenommenen Bild aus der Vielzahl von aufgenommenen Bildern vorgegeben ist, wobei das zweite aufgenommene Bild nach dem ersten aufgenommenen Bild fotografiert wird und vor einem Zeitpunkt fotografiert wird, zu dem das Objekt die eine Rotation ausführt, Positionsinformationen (x, y) bezüglich Positionen des Merkmalspunktes (60) in dem ersten aufgenommenen Bild und dem zweiten aufgenommenen Bild zu erhalten;

Mittel zum Erhalten von Rotationswinkelinformationen (56), die konfiguriert sind, um einen Radius (r) für jeden Merkmalspunkt (60) basierend auf den aufgenommenen Bildern zu berechnen, um Rotationswinkelinformationen (θ1, θ2) bezüglich eines Rotationswinkels des Objekts zwischen einem Zeitpunkt, zu dem das erste aufgenommene Bild fotografiert wird, und einem Zeitpunkt, zu dem das zweite aufgenommene Bild fotografiert wird, basierend auf den Positionsinformationen und dem berechneten Radius zu erhalten; und

Mittel zum Generieren von Bewegtbilddaten (58), um Bewegtbilddaten (42) zu generieren, die zeigen, wie sich das Objekt dreht (30), indem eine Vielzahl von Einzelbildern der Bewegtbilddaten (42) aus der Vielzahl von aufgenommenen Bildern (40) basierend auf den Rotationswinkelinformationen (θ1, θ2), die durch das Mittel zum Erhalten von Rotationswinkelinformationen für jedes der Vielzahl von aufgenommenen Bildern (40) erhalten werden, ausgewählt wird.


 
2. Bildverarbeitungsvorrichtung (10) nach Anspruch 1,
wobei das Mittel zum Erhalten von Rotationswinkelinformationen (56) umfasst:

Mittel zum Erhalten von Positionsdifferenzinformationen (x1, x2) bezüglich einer Differenz zwischen der Position des Merkmalspunktes in dem ersten aufgenommenen Bild und der Position des Merkmalspunktes in dem zweiten aufgenommenen Bild und

Mittel zum Erhalten der Rotationswinkelinformationen (θ1, θ2) basierend auf den Positionsdifferenzinformationen (x1, x2).


 
3. Bildverarbeitungsvorrichtung (10) nach Anspruch 2, umfassend:

Mittel zum Einstellen eines Bereichs in dem aufgenommenen Bild als Beobachtungsbereich (110) basierend auf einer Position einer Rotationsachse (32) des Objekts (30) in dem aufgenommenen Bild,

wobei das Mittel zum Erhalten von Rotationswinkelinformationen (56) konfiguriert ist, um den Merkmalspunkt (60) nur zu verwenden, um die Rotationswinkelinformationen (θ1, θ2) basierend auf den Positionsdifferenzinformationen (x1, x2) zu erhalten, wenn der Merkmalspunkt (60) in dem Beobachtungsbereich (110) enthalten ist.


 
4. Bildverarbeitungsvorrichtung (10) nach Anspruch 2 oder 3,
wobei das Mittel zum Erhalten von Rotationswinkelinformationen (56) umfasst:

Mittel zum Erhalten von Rotationsradiusinformationen, um in einem Fall, in dem eine Vielzahl von Merkmalspunkten (60), die jeweils einer Vielzahl von Punkten an dem Objekt entsprechen, in dem ersten aufgenommenen Bild (40) und dem zweiten aufgenommenen Bild (40) vorgegeben ist, Rotationsradiusinformationen bezüglich eines Rotationsradius eines Punktes an dem Objekt (30), der dem Merkmalspunkt (60) entspricht, für jeden der Vielzahl von Merkmalspunkten (60) zu erhalten;

Mittel zum Erhalten, für jeden der Vielzahl von Merkmalspunkten (60), von Rotationswinkelinformationen bezüglich eines Rotationswinkels des Punktes, der dem Merkmalspunkt (60) an dem Objekt (30) zwischen einem Zeitpunkt, zu dem das erste aufgenommene Bild fotografiert wird, und einem Zeitpunkt, zu dem das zweite aufgenommene Bild fotografiert wird, entspricht, basierend auf den Positionsdifferenzinformationen (x1, x2) und den Rotationsradiusinformationen der Merkmalspunkte; und

Mittel zum Erhalten der Rotationswinkelinformationen des Objekts basierend auf den Rotationswinkelinformationen jedes der Vielzahl von Merkmalspunkten.


 
5. Bildverarbeitungsvorrichtung (10) nach Anspruch 4, wobei das Mittel zum Erhalten von Rotationswinkelinformationen (56) die Rotationswinkelinformationen (θ1, θ2) unter Verwendung der Positionsdifferenzinformationen (x1, x2) eines Merkmalspunktes (60) mit einem ersten Rotationsradius, wenn sich das Objekt (30) dreht, bevorzugt gegenüber den Positionsdifferenzinformationen eines Merkmalspunktes (60) mit einem zweiten Rotationsradius, wenn sich das Objekt (30) dreht, der kleiner als der erste Rotationsradius ist, erhält.
 
6. Bildverarbeitungsvorrichtung (10) nach Anspruch 4,
wobei das Mittel zum Erhalten der Rotationsradiusinformationen umfasst:

Mittel zum Erhalten, in einem Fall, in dem der Merkmalspunkt (60), der dem gewissen Punkt an dem Objekt (30) entspricht, in mindestens drei aufgenommenen Bildern (40) vorgegeben ist, von elliptischen Kreisbahninformationen bezüglich einer elliptischen Kreisbahn des Merkmalspunktes (60) basierend auf der Position des Merkmalspunktes (60) in den mindestens drei aufgenommenen Bildern (40) und

Mittel zum Erhalten der Rotationsradiusinformationen des Merkmalspunktes (60) basierend auf den elliptischen Kreisbahninformationen des Merkmalspunktes (60).


 
7. Bildverarbeitungsvorrichtung (10) nach Anspruch 1,
wobei das Mittel zum Erhalten von Rotationswinkelinformationen (56) umfasst:

Mittel zum Erhalten, in einem Fall, in dem der Merkmalspunkt (60), der dem gewissen Punkt an dem Objekt (30) entspricht, in mindestens drei aufgenommenen Bildern vorgegeben ist, von elliptischen Kreisbahninformationen bezüglich einer elliptischen Kreisbahn des Merkmalspunktes (60) und

Mittel zum Erhalten der Rotationswinkelinformationen basierend auf den Positionen des Merkmalspunktes (60) in dem ersten aufgenommenen Bild und in dem zweiten aufgenommenen Bild und den elliptischen Kreisbahninformationen.


 
8. Bildverarbeitungsvorrichtung (10) nach einem der Ansprüche 1 bis 7,
wobei das Mittel zum Generieren von Bewegtbilddaten (58) umfasst:

Mittel zum Erhalten von Ähnlichkeitsinformationen bezüglich einer Ähnlichkeit zwischen einem aufgenommenen Referenzbild (40), das aus der Vielzahl von aufgenommenen Bildern (40) ausgewählt wird, und jedem von einer Vielzahl von aufgenommenen Bildern (40), die fotografiert werden, nachdem ein vorbestimmter Zeitraum seit einem Zeitpunkt, zu dem das aufgenommene Referenzbild (40) fotografiert wird, vergeht; und

Mittel zum Auswählen eines der Vielzahl von aufgenommenen Bildern (40), die fotografiert werden, nachdem der vorbestimmte Zeitraum seit dem Zeitpunkt, zu dem das aufgenommene Referenzbild (40) fotografiert wird, vergeht, als ein aufgenommenes Bild (40) eines Zeitpunktes, zu dem das Objekt (30) eine Rotation ausführt, basierend auf den Ähnlichkeitsinformationen, und

wobei das Mittel zum Generieren von Bewegtbilddaten (58) die Bewegtbilddaten (42), die zeigen, wie sich das Objekt einmal dreht, basierend auf aufgenommenen Bildern (40), die vor dem Zeitpunkt fotografiert werden, zu dem das aufgenommene Bild (40), das als das aufgenommene Bild (40) des Zeitpunktes, zu dem das Objekt (30) eine Rotation ausführt, ausgewählt wird, generiert.


 
9. Bildverarbeitungsvorrichtung (10) nach Anspruch 8,
wobei das Mittel zum Generieren von Bewegtbilddaten (58) umfasst:

Mittel zum Erhalten einer Anzahl von Merkmalspunkten, um eine Anzahl von Merkmalspunkten (60), die Merkmalspunkten (60) entsprechen, die in dem aufgenommenen Referenzbild (40) vorgegeben sind, aus den Merkmalspunkten (60), die in dem aufgenommenen Bild (40) vorgegeben sind, das fotografiert wird, nachdem der vorbestimmte Zeitraum seit dem Zeitpunkt, zu dem das aufgenommene Referenzbild (40) fotografiert wird, vergeht, zu erhalten;

Mittel zum Auswählen von Kandidaten des aufgenommenen Bildes (40) des Zeitpunktes, zu dem das Objekt (30) eine Rotation ausführt, aus der Vielzahl von aufgenommenen Bildern (40), die fotografiert werden, nachdem der vorbestimmte Zeitraum seit dem Zeitpunkt, zu dem das aufgenommene Referenzbild (40) fotografiert wird, vergeht, basierend auf einem Ergebnis, das durch das Mittel zum Erhalten von Merkmalspunkten erhalten wird;

Mittel zum Erhalten von Ähnlichkeitsinformationen bezüglich einer Ähnlichkeit zwischen dem aufgenommenen Referenzbild (40) und jedem der aufgenommenen Bilder (40), die als die Kandidaten ausgewählt werden; und

Mittel zum Auswählen eines der aufgenommenen Bilder (40), die als die Kandidaten ausgewählt werden, als das aufgenommene Bild (40) eines Zeitpunktes, zu dem das Objekt (30) eine Rotation ausführt, basierend auf den Ähnlichkeitsinformationen.


 
10. Bildverarbeitungsvorrichtung (10) nach einem der Ansprüche 1 bis 9,
wobei die Vielzahl von aufgenommenen Bildern (40) die 1. bis m. (m: Ganzzahl von 2 oder mehr) aufgenommenen Bilder und die m+1. bis m+n. (n: Ganzzahl von 1 oder mehr) aufgenommenen Bilder, die nach dem m. aufgenommenen Bild fotografiert werden, umfasst,
wobei das Mittel zum Generieren von Bewegtbilddaten (58) umfasst:

Mittel zum Erhalten einer Auswahlanzahl, um eine Anzahl der aufgenommenen Bilder (40), die als Einzelbilder der Bewegtbilddaten (42) aus den 1. bis m. aufgenommenen Bildern ausgewählt werden, zu erhalten; und

Mittel zum Auswählen der Einzelbilder der Bewegtbilddaten (42) aus den m+1. bis m+n. aufgenommenen Bildern in einem Fall, in dem der Merkmalspunkt (60), der dem Punkt an dem Objekt (30) entspricht, nicht in den m+1. bis m+n. aufgenommenen Bildern vorgegeben ist, basierend auf einem Verhältnis einer Anzahl der 1. bis m. aufgenommenen Bilder zu der Anzahl der aufgenommenen Bilder (40), die durch das Mittel zum Erhalten einer Auswahlanzahl erhalten werden.


 
11. Bildverarbeitungsvorrichtung (10) nach einem der Ansprüche 1 bis 10, umfassend:

ein erstes Bestimmungsmittel, um zu bestimmen, ob ein Merkmalsbetrag des Merkmalspunktes (60), der in dem ersten aufgenommenen Bild vorgegeben ist, und ein Merkmalsbetrag des Merkmalspunktes (60), der in dem zweiten aufgenommenen Bild vorgegeben ist, gleich oder im Wesentlichen gleich sind oder nicht;

ein zweites Bestimmungsmittel, um zu bestimmen, ob eine Differenz eines Koordinatenwertes einer Richtung einer Achse, die einer Rotationsachse (32) des Objekts (30) zwischen der Position des Merkmalspunktes (60), der in dem ersten aufgenommenen Bild vorgegeben ist, und der Position des Merkmalspunktes (60), der in dem zweiten aufgenommenen Bild vorgegeben ist, entspricht, gleich oder größer als der Schwellenwert ist oder nicht; und

Mittel, um zu bestimmen, ob der Merkmalspunkt (60), der in dem ersten aufgenommenen Bild vorgegeben ist, und der Merkmalspunkt (60), der in dem zweiten aufgenommenen Bild vorgegeben ist, Merkmalspunkte sind oder nicht, die dem gewissen Punkt an dem Objekt (30) entsprechen, basierend auf einem Bestimmungsergebnis des ersten Bestimmungsmittels und des zweiten Bestimmungsmittels.


 
12. Bildverarbeitungsvorrichtung (10) nach einem der Ansprüche 1 bis 11, umfassend:

Bestimmungsmittel, um zu bestimmen, ob der Merkmalsbetrag des Merkmalspunktes (60), der in dem ersten aufgenommenen Bild vorgegeben ist, und der Merkmalsbetrag des Merkmalspunktes (60), der in dem zweiten aufgenommenen Bild vorgegeben ist, gleich oder im Wesentlichen gleich sind oder nicht;

Vergleichsmittel zum Vergleichen von (a) Informationen bezüglich einer Differenz eines Koordinatenwertes einer Richtung einer zweiten Achse, die zu einer ersten Achse, die einer Rotationsachse (32) des Objekts (30) entspricht, senkrecht ist, zwischen der Position eines ersten Merkmalspunktes in dem ersten aufgenommenen Bild und der Position des ersten Merkmalspunktes in dem zweiten aufgenommenen Bild mit (b) Informationen bezüglich einer Differenz eines Koordinatenwertes der Richtung der zweiten Achse zwischen der Position eines zweiten Merkmalspunktes in dem ersten aufgenommenen Bild und der Position des zweiten Merkmalspunktes in dem zweiten aufgenommenen Bild in einem Fall, in dem der erste Merkmalspunkt, der den gleichen oder im Wesentlichen den gleichen Merkmalsbetrag sowohl in dem ersten aufgenommenen Bild als auch in dem zweiten aufgenommenen Bild aufweist, und der zweite Merkmalspunkt, der den gleichen oder im Wesentlichen den gleichen Merkmalsbetrag sowohl in dem ersten aufgenommenen Bild als auch in dem zweiten aufgenommenen Bild aufweist, vorgegeben sind; und

Mittel, um zu bestimmen, ob der Merkmalspunkt, der in dem ersten aufgenommenen Bild vorgegeben ist, und der Merkmalspunkt, der in dem zweiten aufgenommenen Bild vorgegeben ist, Merkmalspunkte (60) sind, die dem gewissen Punkt an dem Objekt (30) entsprechen oder nicht, basierend auf einem Bestimmungsergebnis des Bestimmungsmittels und einem Vergleichsergebnis des Vergleichsmittels.


 
13. Bildverarbeitungsvorrichtung (10) nach einem der Ansprüche 1 bis 12,
wobei das Mittel zum Erhalten aufgenommener Bilder (50) eine Vielzahl von aufgenommenen Bildern (40) erhält, die erhalten werden, indem das Objekt (30) wiederholt fotografiert wird, während sich das Objekt (30) eine Vielzahl von Malen dreht, und
wobei das Mittel zum Generieren von Bewegtbilddaten (58) die Bewegtbilddaten (42), die zeigen, wie sich das Objekt (30) einmal dreht, basierend auf der Vielzahl von aufgenommenen Bildern (40), die erhalten werden, indem das Objekt (30) wiederholt fotografiert wird, während sich das Objekt (30) die Vielzahl von Malen dreht, generiert.
 
14. Bildverarbeitungsverfahren, umfassend:

einen Schritt zum Erhalten eines aufgenommenen Bildes (S102), der darin besteht, aufgenommene Bilddaten einer Vielzahl von aufgenommenen Bildern (40) zu lesen, die erhalten werden, indem ein Objekt (30) wiederholt fotografiert wird, während das Objekt eine Rotation ausführt;

einen Schritt zum Vorgeben von Merkmalspunkten (S103), der darin besteht, unter Verwendung eines Algorithmus eine Vielzahl von Merkmalspunkten (60) in der Vielzahl von aufgenommenen Bildern (40), die fotografiert wurden, während das Objekt eine Rotation ausführte, vorzugeben;

einen Schritt zum Erhalten von Positionsinformationen, der darin besteht, in einem Fall, in dem ein Merkmalspunkt (60), der einem gewissen Punkt an dem Objekt (30) entspricht, in einem ersten aufgenommenen Bild und einem zweiten aufgenommenen Bild aus der Vielzahl von aufgenommenen Bildern vorgegeben ist, wobei das zweite aufgenommene Bild nach dem ersten aufgenommenen Bild fotografiert wird und vor einem Zeitpunkt fotografiert wird, zu dem das Objekt die eine Rotation ausführt, Positionsinformationen (x, y) bezüglich Positionen des Merkmalspunktes (60) in dem ersten aufgenommenen Bild und dem zweiten aufgenommenen Bild zu erhalten;

einen Schritt zum Erhalten von Rotationswinkelinformationen, der darin besteht, einen Radius (r) für jeden Merkmalspunkt basierend auf den aufgenommenen Bildern zu berechnen und Rotationswinkelinformationen (θ1, θ2) bezüglich eines Rotationswinkels des Objekts (30) zwischen einem Zeitpunkt, zu dem das erste aufgenommene Bild fotografiert wird, und einem Zeitpunkt, zu dem das zweite aufgenommene Bild fotografiert wird, basierend auf den Positionsinformationen (x, y) und dem berechneten Radius (r) zu erhalten; und

einen Schritt zum Generieren von Bewegtbilddaten, der darin besteht, Bewegtbilddaten (42) zu generieren, indem eine Vielzahl von Einzelbildern der Bewegtbilddaten (42), die zeigen, wie sich das Objekt (30) dreht, aus der Vielzahl von aufgenommenen Bildern (40) basierend auf den Rotationswinkelinformationen (θ1, θ2), die durch den Schritt zum Erhalten von Rotationswinkelinformationen für jedes der Vielzahl von aufgenommenen Bildern (40) erhalten werden, ausgewählt wird.


 
15. Programm zum Bewirken, dass ein Computer funktioniert als:

Mittel zum Erhalten aufgenommener Bilder (50), um aufgenommene Bilddaten einer Vielzahl von aufgenommenen Bildern, die erhalten werden, indem ein Objekt (30) wiederholt fotografiert wird, während das Objekt eine Rotation ausführt, zu lesen;

Mittel zum Vorgeben von Merkmalspunkten (52), um unter Verwendung eines Algorithmus eine Vielzahl von Merkmalspunkten (60) in der Vielzahl von aufgenommenen Bildern, die fotografiert wurden, während das Objekt die eine Rotation ausführte, vorzugeben,

Mittel zum Erhalten von Positionsinformationen (54), um in einem Fall, in dem ein Merkmalspunkt (60), der einem gewissen Punkt an dem Objekt (30) entspricht, in einem ersten aufgenommenen Bild und einem zweiten aufgenommenen Bild aus der Vielzahl von aufgenommenen Bildern vorgegeben ist, wobei das zweite aufgenommene Bild nach dem ersten aufgenommenen Bild fotografiert wird und vor einem Zeitpunkt fotografiert wird, zu dem das Objekt die eine Rotation ausführt, Positionsinformationen (x, y) bezüglich Positionen des Merkmalspunktes in dem ersten aufgenommenen Bild und dem zweiten aufgenommenen Bild zu erhalten;

Mittel zum Erhalten von Rotationswinkelinformationen (56), das konfiguriert ist, um einen Radius (r) für jeden Merkmalspunkt (60) basierend auf den aufgenommenen Bildern zu berechnen, um Rotationswinkelinformationen (θ1, θ2) bezüglich eines Rotationswinkels des Objekts (30) zwischen einem Zeitpunkt, zu dem das erste aufgenommene Bild fotografiert wird, und einem Zeitpunkt, zu dem das zweite aufgenommene Bild fotografiert wird, basierend auf den Positionsinformationen (x, y) und dem berechneten Radius (r) zu erhalten; und

Mittel zum Generieren von Bewegtbilddaten (58), um Bewegtbilddaten (42) zu generieren, indem eine Vielzahl von Einzelbildern der Bewegtbilddaten (42), die zeigen, wie sich das Objekt (30) dreht, aus der Vielzahl von aufgenommenen Bildern (40) basierend auf den Rotationswinkelinformationen (θ1, θ2), die durch das Mittel (56) zum Erhalten von Rotationswinkelinformationen für jedes der Vielzahl von aufgenommenen Bildern (40) erhalten werden, ausgewählt wird.


 


Revendications

1. Dispositif de traitement d'images (10) comprenant :

des moyens d'obtention d'images capturées (50) pour lire des données d'image capturée d'une pluralité d'images capturées qui sont obtenues par la prise de vue de manière répétée d'un objet (30) alors que l'objet (30) effectue une rotation complète ;

des moyens de spécification de points caractéristiques (52) pour spécifier, en utilisant un algorithme, une pluralité de points caractéristiques (60) dans la pluralité d'images capturées (40) qui sont prises alors que l'objet a effectué la rotation complète,

des moyens d'obtention d'informations de position (54) pour obtenir, dans un cas où un point caractéristique (60) correspondant à un certain point sur l'objet (30) est spécifié dans une première image capturée et une seconde image capturée parmi la pluralité d'images capturées, la seconde image capturée étant prise après la première image capturée et prise avant un point temporel où l'objet effectue la rotation complète, des informations de position (x, y) relatives à des positions du point caractéristique (60) dans la première image capturée et la seconde image capturée ;

des moyens d'obtention d'informations d'angle de rotation (56) configurés pour calculer un rayon (r) pour chaque point caractéristique (60) sur la base des images capturées, pour obtenir des informations d'angle de rotation (θ1, θ2) relatives à un angle de rotation de l'objet entre un point temporel où la première image capturée est prise et un point temporel où la seconde image capturée est prise, sur la base des informations de position et du rayon calculé ; et

des moyens de génération de données d'images en mouvement (58) pour générer des données d'images en mouvement (42) qui montrent comment l'objet tourne (30), en sélectionnant une pluralité d'images de trame des données d'images en mouvement (42) parmi la pluralité d'images capturées (40), sur la base des informations d'angle de rotation (θ1, θ2) obtenues par les moyens d'obtention d'informations d'angle de rotation pour chacune de la pluralité d'images capturées (40).


 
2. Dispositif de traitement d'images (10) selon la revendication 1,
dans lequel les moyens d'obtention d'informations d'angle de rotation (56) comprennent :

des moyens pour obtenir des informations de différence de positions (x1, x2) relatives à une différence entre la position du point caractéristique dans la première image capturée et la position du point caractéristique dans la seconde image capturée ; et

des moyens pour obtenir les informations d'angle de rotation (θ1, θ2) sur la base des informations de différence de positions (x1, x2).


 
3. Dispositif de traitement d'images (10) selon la revendication 2, comprenant :

des moyens pour définir une zone dans l'image capturée comme une zone d'observation (110) sur la base d'une position d'un axe de rotation (32) de l'objet (30) dans l'image capturée,

dans lequel les moyens d'obtention d'informations d'angle de rotation (56) sont configurés pour utiliser seulement le point caractéristique (60) pour obtenir les informations d'angle de rotation (θ1, θ2) sur la base des informations de différence de positions (x1, x2) si le point caractéristique (60) est inclus dans la zone d'observation (110).


 
4. Dispositif de traitement d'images (10) selon la revendication 2 ou 3,
dans lequel les moyens d'obtention d'informations d'angle de rotation (56) comprennent :

des moyens d'obtention d'informations de rayon de rotation pour obtenir, dans un cas où une pluralité de points caractéristiques (60) correspondant respectivement à une pluralité de points sur l'objet sont spécifiés dans la première image capturée (40) et la seconde image capturée (40), des informations de rayon de rotation relatives à un rayon de rotation d'un point sur l'objet (30) correspondant au point caractéristique (60), pour chacun de la pluralité de points caractéristiques (60) ;

des moyens pour obtenir, pour chacun de la pluralité de points caractéristiques (60), des informations d'angle de rotation relatives à un angle de rotation du point correspondant au point caractéristique (60) sur l'objet (30) entre un point temporel où la première image capturée est prise et un point temporel où la seconde image capturée est prise, sur la base des informations de différence de positions (x1, x2) et des informations de rayon de rotation des points caractéristiques ; et

des moyens pour obtenir les informations d'angle de rotation de l'objet sur la base des informations d'angle de rotation de chacun de la pluralité de points caractéristiques.


 
5. Dispositif de traitement d'images (10) selon la revendication 4,
dans lequel les moyens d'obtention d'informations d'angle de rotation (56) obtiennent les informations d'angle de rotation (θ1, θ2) en utilisant les informations de différence de positions (x1, x2) d'un point caractéristique (60) avec un premier rayon de rotation quand l'objet (30) tourne, de préférence aux informations de différence de positions d'un point caractéristique (60) avec un second rayon de rotation quand l'objet (30) tourne qui inférieur au premier rayon de rotation.
 
6. Dispositif de traitement d'images (10) selon la revendication 4,
dans lequel les moyens d'obtention d'informations de rayon de rotation comprennent :

des moyens pour obtenir, dans un cas où le point caractéristique (60) correspondant au certain point sur l'objet (30) est spécifié dans au moins trois images capturées (40), des informations d'orbite elliptique relatives à une orbite elliptique du point caractéristique (60) sur la base de la position du point caractéristique (60) dans les au moins trois images capturées (40) ; et

des moyens pour obtenir les informations de rayon de rotation du point caractéristique (60) sur la base des informations d'orbite elliptique du point caractéristique (60).


 
7. Dispositif de traitement d'images (10) selon la revendication 1,
dans lequel les moyens d'obtention d'informations d'angle de rotation (56) comprennent :

des moyens pour obtenir, dans un cas où le point caractéristique (60) correspondant au certain point sur l'objet (30) est spécifié dans au moins trois images capturées (40), des informations d'orbite elliptique relatives à une orbite elliptique du point caractéristique (60) ; et

des moyens pour obtenir les informations d'angle de rotation sur la base des positions du point caractéristique (60) dans la première image capturée et dans la seconde image capturée et des informations d'orbite elliptique.


 
8. Dispositif de traitement d'images (10) selon l'une quelconque des revendications 1 à 7,
dans lequel les moyens de génération de données d'images en mouvement (58) comprennent :

des moyens pour obtenir des informations de similarité relatives à une similarité entre une image capturée de référence (40) qui est sélectionnée à partir de la pluralité d'images capturées (40) et chacune d'une pluralité d'images capturées (40) qui sont prises après qu'une période de temps prédéterminée s'écoule à partir d'un point temporel où l'image capturée de référence (40) est prise ; et

des moyens pour sélectionner une de la pluralité d'images capturées (40) qui sont prises après que la période de temps prédéterminée s'écoule à partir du point temporel où l'image capturée de référence (40) est prise comme une image capturée (40) d'un point temporel où l'objet (30) effectue une rotation complète sur la base des informations de similarité, et

dans lequel les moyens de génération de données d'images en mouvement (58) génèrent les données d'images en mouvement (42) montrant comment l'objet tourne une fois sur la base d'images capturées (40) qui sont prises avant le point temporel où l'image capturée (40) sélectionnée comme l'image capturée (40) du moment où l'objet (30) effectue une rotation complète est prise.


 
9. Dispositif de traitement d'images (10) selon la revendication 8,
dans lequel les moyens de génération de données d'images en mouvement (58) comprennent :

des moyens d'obtention de nombre de points caractéristiques pour obtenir un nombre de points caractéristiques (60) correspondant à des points caractéristiques (60) spécifiés dans l'image capturée de référence (40), parmi les points caractéristiques (60) spécifiés dans l'image capturée (40) qui est prise après que la période de temps prédéterminée s'écoule à partir du point temporel où l'image capturée de référence (40) est prise ;

des moyens pour sélectionner des candidates de l'image capturée (40) du point temporel où l'objet (30) effectue une rotation complète, parmi la pluralité d'images capturées (40) qui sont prises après que la période de temps prédéterminée s'écoule à partir du point temporel où l'image capturée de référence (40) est prise, sur la base d'un résultat obtenu par les moyens d'obtention de nombre de points caractéristiques ;

des moyens pour obtenir des informations de similarité relatives à une similarité entre l'image capturée de référence (40) et chacune des images capturées (40) sélectionnées comme les candidates ; et

des moyens pour sélectionner une des images capturées (40) sélectionnées comme les candidates comme l'image capturée (40) d'un point temporel où l'objet (30) effectue une rotation complète sur la base des informations de similarité.


 
10. Dispositif de traitement d'images (10) selon l'une quelconque des revendications 1 à 9,
dans lequel la pluralité d'images capturée (40) inclut les 1ère à m-ième (m : un entier de 2 ou plus) images capturées et des m+1-ième à m+n-ième (n : un entier de 1 ou plus) images capturées, qui sont prises après la m-ième image capturée,
dans lequel les moyens de génération de données d'images en mouvement (58) comprennent :

des moyens d'obtention de nombre de sélections pour obtenir un nombre des images capturées (40) sélectionnées comme des images de trame des données d'images en mouvement (42) parmi les 1ère à m-ième images capturées ; et

des moyens pour sélectionner les images de trame des données d'images en mouvement (42) parmi les m+1-ième à m+n-ième images capturées, dans un cas où le point caractéristique (60) correspondant au point sur l'objet (30) n'est pas spécifié dans les m+1-ième à m+n-ième images capturées, sur la base d'un rapport entre un nombre des 1ère à m-ième images capturées et du nombre des images capturées (40) obtenues par les moyens d'obtention de nombre de sélections.


 
11. Dispositif de traitement d'images (10) selon l'une quelconque des revendications 1 à 10, comprenant :

des premiers moyens de détermination pour déterminer si une quantité de caractéristique du point caractéristique (60) spécifié dans la première image capturée et une quantité de caractéristique du point caractéristique (60) spécifié dans la seconde image capturée sont ou non les mêmes ou sensiblement les mêmes ;

des seconds moyens de détermination pour déterminer si une différence d'une valeur de coordonnée d'une direction d'un axe correspondant à un axe de rotation (32) de l'objet (30) entre la position du point caractéristique (60) spécifié dans la première image capturée et la position du point caractéristique (60) spécifié dans la seconde image capturée est ou non égale ou supérieure à la valeur de seuil ; et

des moyens pour déterminer si le point caractéristique (60) spécifié dans la première image capturée et le point caractéristique (60) spécifié dans la seconde image capturée sont ou non des points caractéristiques correspondant au certain point sur l'objet (30) sur la base d'un résultat de détermination des premiers moyens de détermination et des seconds moyens de détermination.


 
12. Dispositif de traitement d'images (10) selon l'une quelconque des revendications 1 à 11, comprenant :

des moyens de détermination pour déterminer si la quantité de caractéristique du point caractéristique (60) spécifié dans la première image capturée et la quantité de caractéristique du point caractéristique (60) spécifié dans la seconde image capturée sont ou non les mêmes ou sensiblement les mêmes ;

des moyens de comparaison pour comparer (a) des informations relatives à une différence d'une valeur de coordonnée d'une direction d'un second axe perpendiculaire à un premier axe correspondant à un axe de rotation (32) de l'objet (30) entre la position d'un premier point caractéristique dans la première image capturée et la position du premier point caractéristique dans la seconde image capturée à (b) des informations relatives à une différence d'une valeur de coordonnée de la direction du second axe entre la position d'un second point caractéristique dans la première image capturée et la position du second point caractéristique dans la seconde image capturée, dans un cas où le premier point caractéristique ayant la même ou sensiblement la même quantité de caractéristique à la fois dans la première image capturée et la seconde image capturée et le second point caractéristique ayant la même ou sensiblement la même quantité de caractéristique à la fois dans la première image capturée et la seconde image capturée sont spécifiés ; et

des moyens pour déterminer si le point caractéristique spécifié dans la première image capturée et le point caractéristique spécifié dans la seconde image capturée sont ou non des points caractéristiques (60) correspondant au certain point sur l'objet (30), sur la base d'un résultat de détermination des moyens de détermination et d'un résultat de comparaison des moyens de comparaison.


 
13. Dispositif de traitement d'images (10) selon l'une quelconque des revendications 1 à 12,
dans lequel les moyens d'obtention d'images capturées (50) obtiennent une pluralité d'images capturées (40) qui sont obtenues par une prise de vue de manière répétée de l'objet (30) alors que l'objet (30) tourne une pluralité de fois, et
dans lequel les moyens de génération de données d'images en mouvement (58) génèrent les données d'images en mouvement (42) qui montrent comment l'objet (30) tourne une fois sur la base de la pluralité d'images capturées (40) qui sont obtenues par la prise de vue de manière répétée de l'objet (30) alors que l'objet (30) tourne la pluralité de fois.
 
14. Procédé de traitement d'images comprenant :

une étape d'obtention d'images capturées (S102) pour lire des données d'image capturée d'une pluralité d'images capturées (40) qui sont obtenues par la prise de vue de manière répétée d'un objet (30) alors que l'objet (30) effectue une rotation complète ;

une étape de spécification de points caractéristiques (S103) pour spécifier, en utilisant un algorithme, une pluralité de points caractéristiques (60) dans la pluralité d'images capturées (40) qui sont prises alors que l'objet a effectué la rotation complète,

une étape d'obtention d'informations de position pour obtenir, dans un cas où un point caractéristique (60) correspondant à un certain point sur l'objet (30) est spécifié dans une première image capturée et une seconde image capturée parmi la pluralité d'images capturées, la seconde image capturée étant prise après la première image capturée et prise avant un point temporel où l'objet effectue la rotation complète, des informations de position (x, y) relatives à des positions du point caractéristique (60) dans la première image capturée et la seconde image capturée ;

une étape d'obtention d'informations d'angle de rotation pour calculer un rayon (r) pour chaque point caractéristique sur la base des images capturées, et pour obtenir des informations d'angle de rotation (θ1, θ2) relatives à un angle de rotation de l'objet (30) entre un point temporel où la première image capturée est prise et un point temporel où la seconde image capturée est prise, sur la base des informations de position (x, y) et du rayon calculé (r) ; et

une étape de génération de données d'images en mouvement pour générer des données d'images en mouvement (42) en sélectionnant une pluralité d'images de trame des données d'images en mouvement (42), qui montrent comment l'objet (30) tourne, parmi la pluralité d'images capturées (40), sur la base des informations d'angle de rotation (θ1, θ2) obtenues par l'étape d'obtention d'informations d'angle de rotation pour chacune de la pluralité d'images capturées (40).


 
15. Programme pour amener un ordinateur à fonctionner comme :

des moyens d'obtention d'images capturées (50) pour lire des données d'image capturée d'une pluralité d'images capturées (40) qui sont obtenues par la prise de vue de manière répétée d'un objet (30) alors que l'objet effectue une rotation complète ;

des moyens de spécification de points caractéristiques (52) pour spécifier, en utilisant un algorithme, une pluralité de points caractéristiques (60) dans la pluralité d'images capturées qui sont prises alors que l'objet a effectué la rotation complète,

des moyens d'obtention d'informations de position (54) pour obtenir, dans un cas où un point caractéristique (60) correspondant à un certain point sur l'objet (30) est spécifié dans une première image capturée et une seconde image capturée parmi la pluralité d'images capturées, la seconde image capturée étant prise après la première image capturée et prise avant un point temporel où l'objet effectue la rotation complète, des informations de position (x, y) relatives à des positions du point caractéristique dans la première image capturée et la seconde image capturée ;

des moyens d'obtention d'informations d'angle de rotation (56) configurés pour calculer un rayon (r) pour chaque point caractéristique (60) sur la base des images capturées, pour obtenir des informations d'angle de rotation (θ1, θ2) relatives à un angle de rotation de l'objet (30) entre un point temporel où la première image capturée est prise et un point temporel où la seconde image capturée est prise, sur la base des informations de position (x, y) et du rayon calculé (r) ; et

des moyens de génération de données d'images en mouvement (58) pour générer des données d'images en mouvement (42) en sélectionnant une pluralité d'images de trame des données d'images en mouvement (42), qui montrent comment l'objet (30) tourne, parmi la pluralité d'images capturées (40), sur la base des informations d'angle de rotation (θ1, θ2) obtenues par les moyens d'obtention d'informations d'angle de rotation (56) pour chacune de la pluralité d'images capturées (40).


 




Drawing



























































Cited references

REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description