(19)
(11)EP 2 869 094 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
20.05.2020 Bulletin 2020/21

(21)Application number: 14190459.9

(22)Date of filing:  27.10.2014
(51)Int. Cl.: 
G01V 5/00  (2006.01)

(54)

Stereoscopic imaging systems and methods

Stereoskopische Abbildungssysteme und Verfahren

Systèmes et procédés d'imagerie stéréoscopique


(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: 29.10.2013 CN 201310521748

(43)Date of publication of application:
06.05.2015 Bulletin 2015/19

(73)Proprietor: Nuctech Company Limited
TongFang Building, Shuangqinglu, Haidian District Beijing 100084 (CN)

(72)Inventors:
  • Chen, Zhiqiang
    100084 Beijing (CN)
  • Tu, Zhuowen
    100084 Beijing (CN)
  • Li, Liang
    100084 Beijing (CN)
  • Zhao, Ziran
    100084 Beijing (CN)
  • Zhang, Duokun
    100084 Beijing (CN)
  • Peng, Zhi
    100084 Beijing (CN)
  • Cui, Jin
    100084 Beijing (CN)
  • Lin, Dong
    100084 Beijing (CN)

(74)Representative: Isarpatent 
Patent- und Rechtsanwälte Behnisch Barth Charles Hassa Peckmann & Partner mbB Friedrichstrasse 31
80801 München
80801 München (DE)


(56)References cited: : 
DE-A1-102004 022 427
US-A1- 2006 170 674
FR-A1- 2 919 780
  
  • J P 01 ET AL: "A new stereoscopic X-ray imaging technique using a single X-ray source: theoretical analysis", NDT&E International, 1 January 1996 (1996-01-01), pages 27-35, XP055174721, Retrieved from the Internet: URL:http://www.sciencedirect.com/science/a rticle/pii/0963869595000364/pdf?md5=08ad24 bc759c7b3a5eb020fc39ddf331&pid=1-s2.0-0963 869595000364-main.pdf [retrieved on 2015-03-09]
  
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


[0001] Embodiments of the present disclosure relates to radiography, and particularly to stereoscopic imaging systems and methods using a single source and multiple detectors.

BACKGROUND



[0002] Safety inspection plays an important role in various fields such as campaigns against terrorism, drug trading, and trafficking. Countries around the world have paid more attention to safety inspection in public sites, and imposed stricter requirements on customs inspection of containers, luggage and the like.

[0003] Currently, radiography systems are mainstream means for safety inspection. A general transmission image produced by linear detectors is a 2-Dimension (2D) image. Such detected image suffers from loss of depth information for an object, and severe overlapping of projections. Thus, people's recognition and identification of the object's shape are affected. J. P. O. Evans et al ("A new stereoscopic X-ray imaging technique using a single X-ray source: theoretical analysis" NDT&E International, 1 January 1996 (1996-01-01), pages 27-35) proposed a stereoscopic X-ray system having a single X-ray source and a pair of linear X-ray detector arrays. FR 2919780 A1 (NUCTECH CO LTD; UNIV TSINGHUA; 6 February 2009 (2009-02-06)) discloses a stereoscopic imaging system having a single X-ray source and a pair of linear detector arrays.

SUMMARY OF THE DISCLOSURE



[0004] Stereoscopic imaging systems and methods are provided in view of one or more problems with the conventional technology.

[0005] In an aspect of the present disclosure, a stereoscopic imaging system according to claim 1 is provided.

[0006] In another aspect of the present disclosure, a method in a stereoscopic imaging system, according to claim 6 is provided.

[0007] Through 3D reconstruction using transmission images collected by a plurality linear detector arrays at certain angles, the lost depth information in the transmission images can be recovered so that the detected object is presented in a stereoscopic manner from different view angles. This facilitates better image analysis.

BRIEF DESCRIPTION OF THE DRAWINGS



[0008] The following figures illustrate implementations of the present invention. The figures and implementations provide some embodiments of the present invention in a nonlimiting and non-exclusive manner, in which:

Fig. 1 shows a top view of an image capturing system according to embodiments of the present disclosure;

Fig. 2 is shows a side view of an image capturing system according to embodiments of the present disclosure;

Fig. 3 is a schematic diagram depicting a process of calculating depth information;

Fig. 4 is a schematic diagram depicting an layout along a linear detector arrays;

Fig. 5 is a schematic diagram depicting a geometric correction principle along the linear detector array;

Fig. 6 shows a schematic diagram of a container truck;

Fig. 7 shows a process of identifying and modeling a container;

Fig. 8 shows a process of modeling a truck body; and

Fig. 9 a schematic diagram showing 3D reconstruction result according to embodiments of the present disclosure.


DETAILED DESCRIPTION



[0009] In the following, particular embodiments of the present invention will be detailed. To be noted, the described embodiments are just intended for illustrating other than limiting the present invention. Numerous specific details are illustrated for a clear and thorough understanding of the present invention. It is apparent to those skilled in the art that these specific details are not necessary for implementation of the present invention. Detailed description of known structures, circuits, materials or methods are omitted which otherwise may obscure the present invention.

[0010] Throughout the specification, reference to "an embodiment," "embodiments," "an example" or "examples" means that particular features, structures or characteristics described in connection with such embodiment or example are contained in at least one embodiment of the present invention. The phrase "an embodiment," "embodiments," "an example" or "examples" in various places throughout the specification does not necessarily refer to the same embodiment or example. Further, the particular features, structures or characteristics may be contained in one or more embodiments or examples in any appropriate combination and/or sub-combination. Those skilled in the art will appreciate that the term "and/or" herein indicates any or all combinations of one or more of the listed items.

[0011] 3D images are widely used in people's daily lives. Compared with 2D images, 3D images can better depict real scenes and achieve more vivid visual effects. In the radiography field, CT imaging can reconstruct a 3D structure of an object. However, 3D CT imaging has disadvantages that scanning devices have very complex structure, measurement requires a long time, and so on. In addition to scientific contribution, stereoscopic imaging technology from multiple view angles in transmission images has good application prospect. The stereoscopic imaging technology from multiple view angles in transmission images can give an approximate 3D image. When presented with small-angle rotation or on a 3D display terminal, such 3D image will provide an inspector with a stereoscopic perception, and enhance his or her experience.

[0012] In view of the problems described above, with embodiments of the present disclosure, through 3D reconstruction using transmission images collected by a plurality of linear detector arrays at certain angles, the lost depth information in the transmission images can be recovered so that the detected object is presented in a stereoscopic manner from different view angles. This facilitates better image analysis. For example, in safety inspection of containers, the container may be scanned with X rays without opening the container. Then, 3D information of part of the container may be restored for presentation in a new manner so that a user has a new viewing experience.

[0013] According to embodiments of the present disclosure, a stereoscopic imaging system is provided including a ray source, a plurality of linear detector arrays, and a reconstruction apparatus implemented by a computer, for example. The ray source is configured to emit a plurality of ray fanbeams. Each linear detector array is arranged at a preset angle with respect to the ray source, and configured to detect a strength value of a respective one of the ray fanbeams penetrating an object under inspection and form a respective transmission image, when the object moves along a direction intersecting with the ray fanbeams. The reconstruction apparatus is configured to use any two of the formed transmission images as a binocular image, calculate depth information of the object on the transmission images, superpose and fuse the calculated depth information to obtain 3D information, and perform 3D reconstruction

[0014] Fig. 1 shows a top view of an image capturing system. There are three viewing angles shown in the figure. This is exemplary, and more viewing angles may be used according to requirements of practical applications. Ray fanbeams from the ray sournce 110 to the three (i.e., left, middle, and right) linear detector arrays have a preset beam angle between them, such as an angle θ between the ray fanbeams emitted to the left and middle linear detector arrays, and the same angle θ between the ray fanbeams emitted to the right and middle linear detector arrays. Three images scanned at different angles are then obtained. The middle ray fanbeam is along a main beam direction, and the left and right ray fanbeams are symmetrically located at both sides of the main beam. Although three viewing angles are used in this embodiment, the image capturing system may use more viewing angles by adding pairs of viewing angles at both sides of the main beam in a symmetric manner.

[0015] Fig. 2 shows a side view of an image capturing system. The capturing system primarily includes the ray source 110, an collimation apparatus (not shown), and a data collection system (not shown). When the object under inspection passes a detection area at a velocity, a number of images will be generated simultaneously, and the number of images are equal to the number of linear detector arrays. An appropriate reference coordinate system may be establish in this imaging model to facilitate 3D reconstruction and multi-angle viewing of the obtained images at multiple viewing angles.

[0016] Fig. 3 shows the principle for calculating depth information by using two images at different viewing angles. There will be a difference between positions of objects having different depths in the two images, and the depth information can be acquired with such position difference. Assume that there are two points A and B at different depths, and the point B is on the right of the point A in the right detector-scanned image, while the point B is on the left of the point A in the left detector-scanned image. The relative position difference between the two points is L, and the depth difference between the two points A and B may be calculated with L and a beam angle α as follows:



[0017] When the beam angle α is very small,



[0018] i.e., the relative depth between the objects is in proportion to the relative position difference between the objects in the left and right images.

[0019] The depth information of the objects on any two of the images captured at different viewing angles can be calculated by combining the two images. The calculated 3D information may be superposed and fused to restore 3D information as much as possible. In the example using three viewing angles, three images can be captured at the same time as left, middle and right images. There are three combinations of any two of the images. In the case of combining the left and right images, the coordinate system of one of the two images may be used as a reference coordinate system, and 3D coordinates may be calculated with the following formula:



[0020] Here, v represents the velocity at which the object passes the scan area, f represents a frequency at which rays are emitted, α is an angle sandwiched between the two viewing angles of the two images, and Δx represents a difference between positions of a single feature point in the different images. L denotes a distance from the X-ray source to a plane in which the detectors are arranged. With the left image being a reference image, the x and y coordinates may be modified using the calculated depth information z as follows:





[0021] Here, xi, yi are coordinates of a point in the images, and dy represents a resolution in the y direction. Calculation methods similar to the above may be used for the other combinations of images, except that the formulas for coordinate calculation and modification need to be appropriately adapted.

[0022] Further, there may be geometric distortion along the direction of linear detector array, and it is thus necessary to perform geometric correction in the direction of linear detector array. Causes and correction methods of such geometric distortion will be detailed below.

[0023] Parameters such as device position and beam angle are relatively fixed during the image capturing process, and thus it is unnecessary to consider a process similar to a calibration process for an optical image. Selection of suitable reference coordinate system will suffice.

[0024] According to some embodiments of the present disclosure, point-to-point matching between two images captured at different viewing angles may be obtained by: first establishing a correspondence between a feature point in one of the images and the corresponding point in the other image, by means of conventional methods for feature extraction and matching; then establishing Hilbert Kernels centered on the feature points, and a Hilbert Space by superposition; and iteratively calculating a smooth spatial vector field by using a spatial smoothness constraint. This spatial vector field represents the point-to-point correspondence between the two images.

[0025] Alternatively, the algorithm SIFT Flow may be used to obtain the point-to-point correspondence between two images. The principle of SIFT Flow is first calculating a SIFT eigenvector, Dense SIFT, for each pixel in the images, and then matching a description of the SIFT vectors using a flow estimation algorithm. Unfortunately, the algorithm SIFT Flow equally treats all pixels in the images to be matched, thereby ignoring difference between pixels. each pixel in an image may have a different amount of information. Salient pixels for an object in the image generally have a larger amount of information, while pixels in background (e.g., a background in pure black) have a smaller amount of information. If all the pixels are treated equally, matching values of pixels containing no or little information may influence matching values of important pixels. This incurs disturbance in matching process, and leads to an overall inaccurate result. The present disclosure, however, employs an improved energy function which can increase contribution of areas having a larger amount of information, and weaken contribution of background areas. Accordingly, it is possible to prevent the match process from being influenced by the background area, and facilitate obtaining of more accurate matching result. Moreover, the algorithm Belief Propagation may be used to optimize a target function to calculate an optimized flow field, which represents a correspondence between two images.

[0026] Since X rays are emitted in a radial pattern, the linear detectors may be arranged in a fan(or arc)-shape layout to accommodate the emission pattern of X rays and thus reduce geometric distortion. It is an optimal design to arrange the detectors in a fan(or arc)-shape layout. In practical applications, however, only a layout approximate to a fan(or arc) shape can be achieved due to mechanical architecture and available space. Fig. 4 shows an example of the detector layout. The detectors may be installed linearly on either a vertical arm support or a horizontal arm support, or both of them, i.e., L-shape arm support. For example, a L-shape linear detector array includes a plurality of detector modules 131, 132, and 133 for receiving ray beams emitted from the ray source 110. The angle between the two arm supports can be adjusted. Although the embodiment employs the L-shape arm support, any other suitable layout may be used according to requirements of mechanical architecture in practical application. Such layout approximate to a fan(or arc) shape results in a change in radius of the fan shape, and thus the geometric distortion in the direction of linear detector array cannot be eliminated but reduced. This geometric distortion will severely affect 3D reconstruction, and it is thus necessary to perform geometric correction in the reconstruction process.

[0027] Fig. 5 shows the principle of geometric correction. Given a size S in the direction linear detector array, and a number N of detectors, the resolution in the direction of X-ray main beam is S/N, which is considered as a standard resolution. A distance from the ray source O to the detector modules in the main-beam direction is D, that is, the fan radius is D. For any of the other modules, the distance varies by ΔD. For each of these different modules, ΔD can be obtained based on the layout of the detectors, the size S1 with respect to the module in the main-beam direction is


and the resolution is



[0028] Based on difference in resolution, geometric correction in the direction of linear detector array can be performed using a method of image differencing.

[0029] Fig. 6 shows a container truck under inspection, and is one of the three images captured in the above embodiment.

[0030] Goods in the container are of interest. The container itself is a regular cube. The 3D information of the container itself can be restored perfectly by performing 3D modeling. The obtained 3D information may be used as a reference for 3D reconstruction of the goods within the container. Fig. 7 shows a process of position identification and modeling of the container as explained below.

[0031] At step S110, influence caused by instability of dosage is eliminated.

[0032] The influence on the algorithm from instability of dosage is eliminated by using prior information.



[0033] Here, α represents an air value for a current linear detector array which is capturing an image (i.e., an image captured when there is only air in the relevant area), and I represents a reading of a detector in the current linear detector array. The transformed image gray value Ux is insensitive to instability of dosage.

[0034] A top edge is detected at step S120.

[0035] In an example, the top edge of the container may be detected using a morphological method to calculate a morphological gradient, and using the Hough transform. If no top edge is detected, an error is returned at step S150.

[0036] At step S140, the head part of the truck is removed, leaving only the container part.

[0037] The upper edge of the chassis in the truck is detected in the same way as detection of the top edge of the container.

[0038] Then, the exact position and size of the container can be obtained after detection of the container top edge and the chassis upper edge.

[0039] At step S160, the container is modeled.

[0040] By using the obtained position and size of the container, the type of the container is determined, and accordingly a suitable 3D model is selected from a template library, which will be imported in 3D reconstruction.

[0041] For a container truck under safety inspection, the truck part is generally of less importance. 3D information of the truck part may be restored quickly and effectively by using a modeling method. The position and size of the truck part are first identified as features, and then a best 3D template for the truck part is found in a template library.

[0042] Fig. 8 shows a process of identifying and modeling the truck part. In the present disclosure, the position of the truck part in an image is identified mainly based on edge information. Given that the container has been identified, the type of the truck and the position of the truck head are identified. To obtain a robust result, detection of the truck position may be performed in three steps: pre-processing of strip removal to obtain an image with a gentle change in background; calculating gradients and quantizing a map of the gradients to remove influence from small gradient fluctuations; and locating an maximal continuous area (i.e., the truck position) of horizontal and vertical projections in the binarized gradient map.

[0043] At step S210, the pre-processing of strip removal includes removing strips in each of horizontal and vertical directions. In the horizontal direction, a sequence of projections of the image in the vertical direction is first acquired. Each projection is subjected to median filtering, and the image line corresponding to the projection is determined as a stripe if there is a large difference before and after the filtering. Values in the image line are replaced with those of the closest, non-stripe image line.

[0044] At step S220, the gradients are obtained by quantizing the image and then calculating the gradients.

[0045] The truck position is obtained by calculating projections in horizontal and vertical directions with respect to the gradient map, and detecting a maximal continuous area after subtracting the minimal value (i.e., removing any stripe which possibly remains). This detected area represents the truck position.

[0046] At step S230, the container is identified as above described.

[0047] At step S240, with the container having been identified, the position and orientation of the truck are determined by identifying the truck part. The 3D model of the truck from the template library can be used as the truck part during 3D reconstruction.

[0048] In the safety inspection industry, stereoscopic reconstruction from transmission images is for the purpose of assisting in inspection, and accurate measurement is not required. Accordingly, the present disclosure focuses on rendering a stereoscopic visualization of transmission images by using depth information contained in the transmission images. The purpose is to distinguishably display the images of objects overlapped with each other. The stereoscopic visualization may be presented in two approaches. The first approach is to use a 3D presentation tool, such as OpenGL, to develop 3D presentation software which provides a 3D effect by rotation at different angles. The second approach is using a hardware presentation terminal. In the present disclosure, a 3D display for naked eye is used to display 3D data. Fig. 9 depicts a result of 3D reconstruction presented via OpenGL software kit.

[0049] With the above embodiments, a method for assisting in inspection is provided by rendering a stereoscopic visualization of transmission images at different viewing angles based on depth information calculated from the transmission images. Further, the 3D modeling of the container is implemented with modeling techniques to improve accuracy in 3D reconstruction. Third, the 3D modeling of the truck is implemented with modeling techniques to enhance 3D presentation.

[0050] In addition, the 3D presentation can be provided at different angles by using presentation software. The 3D presentation can be further enhanced in conjunction with hardware presentation terminal.

[0051] The present disclosure has been described with reference to several exemplary embodiments. It will be appreciated that the terms used here are for illustration, are exemplary other than limiting. The present disclosure can be practiced in various forms within the subject matter of the present disclosure. It will be appreciated that the foregoing embodiments are not limited to any of the above detailed description, and should be construed in a broad sense within the scope defined by the appended claims. All changes and variations falling into the scope of the claims should be encompassed by the appended claims.


Claims

1. A stereoscopic imaging system, comprising:

an X-ray source (110) configured to emit a plurality of x-ray fanbeams;

at least three linear detector arrays (120, 130, 140), wherein each linear detector array is arranged at a preset angle with respect to the x-ray source, and configured to detect a strength value of a respective one of the x-ray fanbeams penetrating an object under inspection and form a respective transmission image, when the object moves along a direction intersecting with the x-ray fanbeams; and

a reconstruction apparatus configured to use any two of the formed transmission images as a binocular image, calculate depth information of the object on the transmission images, superpose and fuse the calculated depth information to obtain 3-Dimensional (3D) information, and perform 3D reconstruction.


 
2. The stereoscopic imaging system according to claim 1, wherein the reconstruction apparatus is configured to calculate the depth information based on difference in positions of the object of different depths in the two images.
 
3. The stereoscopic imaging system according to claim 1, wherein the reconstruction apparatus is configured to correct x and y coordinates using the calculated depth information z.
 
4. The stereoscopic imaging system according to claim 1, wherein the reconstruction apparatus is configured to calculate the depth information by
assuming that there are two points A and B of different depths, the point B is on the right of the point A in a right detector scan image, and the point B is on the left of the point A in a left detector scan image, and a relative positional difference between the two points is L, and calculating a depth difference H between the two points A and B with L and a beam angle α as


 
5. The stereoscopic imaging system according to claim 1, wherein the reconstruction apparatus is configured to identify a container, determine a position of the container and perform a 3D reconstruction of the container, based on one of the transmission images.
 
6. A method in a stereoscopic imaging system comprising an X-ray ray source and at least three linear detector arrays, wherein each linear detector array is arranged at a preset angle with respect to the x-ray source, the method comprises:

emitting a plurality of x-ray fanbeams;

detecting, by each linear detector array, a strength value of a respective one of the x-ray fanbeams penetrating an object under inspection and forming a respective transmission image, when the object moves along a direction intersecting with the x-ray fanbeams; and

using any two of the formed transmission images as a binocular image, calculating depth information of the object on the transmission images, superposing and fusing the calculated depth information to obtain 3-Dimensional (3D) information, and performing 3D reconstruction.


 
7. The method according to claim 6, wherein the depth information is calculated based on difference in positions of the object of different depths in the two images.
 
8. The method according to claim 6, wherein the calculated depth information z is used to correct x and y coordinates.
 
9. The method according to claim 6, wherein the depth information is calculated by assuming that there are two points A and B of different depths, the point B is on the right of the point A in a right detector scan image, and the point B is on the left of the point A in a left detector scan image, and a relative positional difference between the two points is L, and calculating a depth difference H between the two points A and B with L and a beam angle α as


 


Ansprüche

1. Stereoskopisches Bildgabesystem, umfassend:

eine Röntgenquelle (110), die dafür eingerichtet ist, mehrere Röntgenfächerstrahlen auszusenden;

mindestens drei lineare Detektoranordnungen (120, 130, 140), wobei jede lineare Detektoranordnung in einem voreingestellten Winkel in Bezug auf die Röntgenquelle angeordnet und dafür eingerichtet ist, einen Stärkewert eines jeweiligen der Röntgenfächerstrahlen zu detektieren, die ein zu inspizierendes Objekt durchdringen, und ein jeweiliges Übertragungsbild zu bilden, wenn sich das Objekt entlang einer Richtung bewegt, die sich mit den Röntgenfächerstrahlen schneidet; und

eine Rekonstruktionsvorrichtung, die dafür eingerichtet ist, jeweils zwei der gebildeten Übertragungsbilder als ein Binokularbild zu verwenden, Tiefeninformationen des Objekts auf den Übertragungsbildern zu berechnen, die berechneten Tiefeninformationen zu überlagern und zu fusionieren, um 3-dimensionale (3D-) Informationen zu erhalten, und eine 3D-Rekonstruktion auszuführen.


 
2. Stereoskopisches Bildgabesystem nach Anspruch 1, wobei die Rekonstruktionsvorrichtung dafür eingerichtet ist, die Tiefeninformationen auf der Grundlage einer Differenz der Positionen des Objekts unterschiedlicher Tiefen in den beiden Bildern zu berechnen.
 
3. Stereoskopisches Bildgabesystem nach Anspruch 1, wobei die Rekonstruktionsvorrichtung dafür eingerichtet ist, x- und y-Koordinaten unter Verwendung der berechneten Tiefeninformation z zu korrigieren.
 
4. Stereoskopisches Bildgabesystem nach Anspruch 1, wobei die Rekonstruktionsvorrichtung dafür eingerichtet ist, die Tiefeninformationen zu berechnen, indem angenommen wird, dass es zwei Punkte A und B unterschiedlicher Tiefen gibt, wobei sich Punkt B rechts von Punkt A in einem rechten Detektorabtastbild befindet und Punkt B sich links von Punkt A in einem linken Detektorabtastbild befindet, und eine relative Positionsdifferenz zwischen den beiden Punkten L ist, und eine Tiefendifferenz H zwischen den beiden Punkten A und B mit L und einem Strahlwinkel α berechnet wird als


 
5. Stereoskopisches Bildgabesystem nach Anspruch 1, wobei die Rekonstruktionsvorrichtung dafür eingerichtet ist, einen Container zu identifizieren, eine Position des Containers zu bestimmen und eine 3D-Rekonstruktion des Containers auf der Grundlage eines der Übertragungsbilder auszuführen.
 
6. Verfahren in einem stereoskopischen Bildgabesystem, das eine Röntgenstrahlquelle und mindestens drei lineare Detektoranordnungen umfasst, wobei jede lineare Detektoranordnung in einem vorgegebenen Winkel in Bezug auf die Röntgenstrahlquelle angeordnet ist, wobei das Verfahren umfasst:

Aussenden mehrerer Röntgenfächerstrahlen;

Detektieren, durch jede lineare Detektoranordnung, eines Stärkewertes eines jeweiligen der Röntgenfächerstrahlen, die ein zu inspizierendes Objekt durchdringen, und Bilden eines jeweiligen Übertragungsbildes, wenn sich das Objekt entlang einer Richtung bewegt, die sich mit den Röntgenfächerstrahlen schneidet; und

Verwenden von jeweils zwei der gebildeten Übertragungsbilder als ein Binokularbild, Berechnen von Tiefeninformationen des Objekts auf den Übertragungsbildern, Überlagern und Fusionieren der berechneten Tiefeninformationen, um 3-dimensionale (3D-) Informationen zu erhalten, und Ausführen einer 3D-Rekonstruktion.


 
7. Verfahren nach Anspruch 6, wobei die Tiefeninformationen auf der Grundlage einer Differenz der Positionen des Objekts unterschiedlicher Tiefen in den beiden Bildern berechnet werden.
 
8. Verfahren nach Anspruch 6, wobei die berechnete Tiefeninformation z zum Korrigieren von x- und y-Koordinaten verwendet wird.
 
9. Verfahren nach Anspruch 6, wobei die Tiefeninformationen berechnet werden, indem angenommen wird, dass es zwei Punkte A und B unterschiedlicher Tiefen gibt, wobei sich Punkt B rechts von Punkt A in einem rechten Detektorabtastbild befindet und Punkt B sich links von Punkt A in einem linken Detektorabtastbild befindet, und eine relative Positionsdifferenz zwischen den beiden Punkten L ist, und eine Tiefendifferenz H zwischen den beiden Punkten A und B mit L und einem Strahlwinkel α berechnet wird als


 


Revendications

1. Système d'imagerie stéréoscopique comprenant :

une source de rayons X (110) conçue pour émettre une pluralité de faisceaux de rayons X en éventail ;

au moins trois groupements de détecteurs linéaires (120, 130, 140),

dans lequel chaque groupement de détecteurs linéaires est disposé à un angle prédéfini par rapport à la source de rayons X, et est conçu pour détecter une valeur d'intensité de l'un respectif des faisceaux de rayons X en éventail pénétrant un objet en cours d'examen et pour former une image de transmission respective, lorsque l'objet se déplace le long d'une direction croisant les faisceaux de rayons X en éventail ; et

un appareil de reconstruction conçu pour utiliser deux quelconques des images de transmission formées, en tant qu'image binoculaire, calculer des informations de profondeur de l'objet sur les images de transmission,

superposer et fusionner les informations de profondeur calculées pour obtenir des informations tridimensionnelles (en 3D), et réaliser une reconstruction en 3D.


 
2. Système d'imagerie stéréoscopique selon la revendication 1, dans lequel l'appareil de reconstruction est conçu pour calculer les informations de profondeur sur la base d'une différence de positions de l'objet de profondeurs différentes sur les deux images.
 
3. Système d'imagerie stéréoscopique selon la revendication 1, dans lequel l'appareil de reconstruction est conçu pour corriger des coordonnées x et y à l'aide des informations de profondeur calculées z.
 
4. Système d'imagerie stéréoscopique selon la revendication 1, dans lequel l'appareil de reconstruction est conçu pour calculer les informations de profondeur en supposant qu'il existe deux points A et B de profondeurs différentes, que le point B est à droite du point A sur une image droite de balayage de détecteur, et que le point B est à gauche du point A sur une image gauche de balayage de détecteur, et qu'une différence de position relative entre les deux points est L, et en calculant une différence de profondeur H entre les deux points A et B avec L et un angle de faisceau a, de la manière suivante :


 
5. Système d'imagerie stéréoscopique selon la revendication 1, dans lequel l'appareil de reconstruction est conçu pour identifier un conteneur, déterminer une position du conteneur et réaliser une reconstruction en 3D du conteneur, sur la base de l'une des images de transmission.
 
6. Procédé dans un système d'imagerie stéréoscopique comprenant une source de rayonnement à rayons X et au moins trois groupements de détecteurs linéaires, dans lequel chaque groupement de détecteurs linéaires est disposé à un angle prédéfini par rapport à la source de rayons X, le procédé comprenant les étapes suivantes :

émettre une pluralité de faisceaux de rayons X en éventail ; détecter, au moyen de chaque groupement de détecteurs linéaires, une valeur d'intensité de l'un respectif des faisceaux de rayons X en éventail pénétrant un objet en cours d'examen et former une image de transmission respective, lorsque l'objet se déplace le long d'une direction croisant les faisceaux de rayons X en éventail ; et

utiliser deux quelconques des images de transmission formées, en tant qu'image binoculaire, calculer des informations de profondeur de l'objet sur les images de transmission, superposer et fusionner les informations de profondeur calculées pour obtenir des informations tridimensionnelles (en 3D), et réaliser une reconstruction en 3D.


 
7. Procédé selon la revendication 6, dans lequel les informations de profondeur sont calculées sur la base d'une différence de positions de l'objet de profondeurs différentes sur les deux images.
 
8. Procédé selon la revendication 6, dans lequel les informations de profondeur calculées z sont utilisées pour corriger des coordonnées x et y.
 
9. Procédé selon la revendication 6, dans lequel les informations de profondeur sont calculées en supposant qu'il existe deux points A et B de profondeurs différentes, que le point B est à droite du point A sur une image droite de balayage de détecteur, et que le point B est à gauche du point A sur une image gauche de balayage de détecteur, et qu'une différence de position relative entre les deux points est L, et en calculant une différence de profondeur H entre les deux points A et B avec L et un angle de faisceau a, de la manière suivante :


 




Drawing















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