(19)
(11)EP 3 644 229 A1

(12)EUROPEAN PATENT APPLICATION

(43)Date of publication:
29.04.2020 Bulletin 2020/18

(21)Application number: 19204291.9

(22)Date of filing:  21.10.2019
(51)International Patent Classification (IPC): 
G06K 9/00(2006.01)
(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
Designated Extension States:
BA ME
Designated Validation States:
KH MA MD TN

(30)Priority: 26.10.2018 US 201816172256

(71)Applicant: HERE Global B.V.
5611 ZT Eindhoven (NL)

(72)Inventors:
  • Pojman, Nicholas
    San Francisco, CA 94109 (US)
  • Mittal, Anish
    San Francisco, CA 94131 (US)
  • Chen, Zhanwei
    Richmond, CA 94801 (US)
  • Lawlor,, David
    Chicago, IL 60601 (US)

(74)Representative: Potter Clarkson 
The Belgrave Centre Talbot Street
Nottingham NG1 5GG
Nottingham NG1 5GG (GB)

  


(54)METHOD, APPARATUS, AND SYSTEM FOR DETERMINING A GROUND CONTROL POINT FROM IMAGE DATA USING MACHINE LEARNING


(57) An approach is provided for determining a ground control point from image data using machine learning. The approach, for example, involves selecting an feature based determining that the feature meets one or more properties for classification as a machine learnable feature. The approach also involves retrieving a plurality of ground truth images depicting the feature. The plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images. The approach further involves training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.




Description

BACKGROUND



[0001] Modern location-based services and applications (e.g., autonomous driving) are increasingly demanding highly accurate and detailed digital map data (e.g., centimeter-level accuracy or better). To achieve such levels of accuracy, map service providers have traditionally used ground control points to precisely align and/or correct digital map data from different sources. Ground control points, for instance, are identifiable points on the Earth's surface that have precise three-dimensional location (e.g., latitude, longitude, and elevation). Traditionally, generating ground control points has been a manual effort that requires deploying ground surveyors to the locations of ground control points to make manual measurements. This traditional approach, however, is labor intensive can does not scale well when available manual resources are limited. Accordingly, map service providers face significant technical challenges to automatically designating and measuring ground control points at the target levels of accuracy.

SOME EXAMPLE EMBODIMENTS



[0002] Therefore, there is a need for an approach for automatically determining ground control points from image data using machine learning.

[0003] According to one embodiment, a computer-implemented method for determining a ground control point from image data comprises selecting a feature based determining that the feature meets one or more properties for classification as a machine learnable feature. The method also comprises retrieving a plurality of ground truth images depicting the feature. The plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images. The method further comprises training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

[0004] According to another embodiment, an apparatus for determining a ground control point from image data comprises at least one processor, and at least one memory including computer program code for one or more computer programs, the at least one memory and the computer program code configured to, with the at least one processor, cause, at least in part, the apparatus to select a feature based determining that the feature meets one or more properties for classification as a machine learnable feature. The apparatus is also caused to retrieve a plurality of ground truth images depicting the feature. The plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images. The apparatus is further caused to train a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

[0005] According to another embodiment, a non-transitory computer-readable storage medium for determining a ground control point from image data carries one or more sequences of one or more instructions which, when executed by one or more processors, cause, at least in part, an apparatus to select a feature based determining that the feature meets one or more properties for classification as a machine learnable feature. The apparatus is also caused to retrieve a plurality of ground truth images depicting the feature. The plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images. The apparatus is further caused to train a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

[0006] According to another embodiment, an apparatus for determining a ground control point from image data comprises means for selecting a feature based determining that the feature meets one or more properties for classification as a machine learnable feature. The apparatus also comprises means for retrieving a plurality of ground truth images depicting the feature. The plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images. The apparatus further comprises means for training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

[0007] The following numbered paragraphs are also disclosed:
  1. 1. A computer-implemented method for determining a ground control point from image data comprising:

    selecting a feature as the ground control point based determining that the feature meets one or more properties for classification as a machine learnable feature;

    retrieving a plurality of ground truth images depicting the feature, wherein the plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images; and

    training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

  2. 2. The method of paragraph 1, wherein the feature is selected based on determining that the one or more properties indicate that the feature is uniquely identifiable from among other features.
  3. 3. The method of paragraph 1, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature has a spatial sparsity that meets a sparsity criterion.
  4. 4. The method of paragraph 1, wherein the feature is selected based on determining that the one or more properties indicate that the feature is applicable to a plurality of different geographic regions.
  5. 5. The method of paragraph 1, wherein the feature is selected from a category of curvilinear geometry features intersection features including a gore point, a cross-walk corner, a boundary of a land boundary and a limit line, or a combination thereof.
  6. 6. The method of paragraph 1, wherein feature is a visible feature of a roadway intersection that is visible from a top-down imagery perspective.
  7. 7. The method of paragraph 1, wherein the training features for the machine learning model include a pixel location of the intersection feature in said each of the plurality of ground truth images, a camera position of said each of the plurality of ground truth images, an orientation of said each of the plurality of ground truth images, a focal length of said each of the plurality of ground truth images, or a combination thereof.
  8. 8. The method of paragraph 1, further comprising:

    processing a plurality of input images using the trained machine learning model to identify that the ground control point is present in the plurality of input images; and

    triangulating a real-world location of the ground control point based on a pixel correspondence of the ground control point among the plurality of input images.

  9. 9. The method of paragraph 8, wherein the real-world location is a three-dimensional location of the ground control point.
  10. 10. The method of paragraph 1, wherein the machine learning model is further trained to calculate an uncertainty associated with the predicted location based on a characteristic of said each of the plurality of images, a respective source of said each of the plurality of images, or a combination.
  11. 11. An apparatus for determining a ground control point from image data comprising:

    at least one processor; and

    at least one memory including computer program code for one or more programs,

    the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following,

    select a feature as the ground control point based determining that the feature meets one or more properties for classification as a machine learnable feature;

    retrieve a plurality of ground truth images depicting the feature, wherein the plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images; and

    train a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

  12. 12. The apparatus of paragraph 11, wherein the feature is selected based on determining that the one or more properties indicate that the feature is uniquely identifiable from among other features.
  13. 13. The apparatus of paragraph 11, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature has a spatial sparsity that meets a sparsity criterion.
  14. 14. The apparatus of paragraph 11, wherein the feature is selected based on determining that the one or more properties indicate that the feature is applicable to a plurality of different geographic regions.
  15. 15. The apparatus of paragraph 11, wherein the feature is selected from a category of curvilinear geometry features intersection features including a gore point, a cross-walk corner, a boundary of a land boundary and a limit line, or a combination thereof.
  16. 16. A non-transitory computer-readable storage medium for determining a ground control point from image data, carrying one or more sequences of one or more instructions which, when executed by one or more processors, cause an apparatus to perform:

    selecting a feature as the ground control point based determining that the feature meets one or more properties for classification as a machine learnable feature;

    retrieving a plurality of ground truth images depicting the feature, wherein the plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images; and

    training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.

  17. 17. The non-transitory computer-readable storage medium of paragraph 16, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature is uniquely identifiable from among other intersection features.
  18. 18. The non-transitory computer-readable storage medium of paragraph 16, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature has a spatial sparsity that meets a sparsity criterion.
  19. 19. The non-transitory computer-readable storage medium of paragraph 16, wherein the feature is selected based on determining that the one or more properties indicate that the feature is applicable to a plurality of different geographic regions.
  20. 20. The non-transitory computer-readable storage medium of paragraph 16, wherein the feature is selected from a category of curvilinear geometry features intersection features including a gore point, a cross-walk corner, a boundary of a land boundary and a limit line, or a combination thereof.


[0008] In addition, for various example embodiments of the invention, the following is applicable: a method comprising facilitating a processing of and/or processing (1) data and/or (2) information and/or (3) at least one signal, the (1) data and/or (2) information and/or (3) at least one signal based, at least in part, on (or derived at least in part from) any one or any combination of methods (or processes) disclosed in this application as relevant to any embodiment of the invention.

[0009] For various example embodiments of the invention, the following is also applicable: a method comprising facilitating access to at least one interface configured to allow access to at least one service, the at least one service configured to perform any one or any combination of network or service provider methods (or processes) disclosed in this application.

[0010] For various example embodiments of the invention, the following is also applicable: a method comprising facilitating creating and/or facilitating modifying (1) at least one device user interface element and/or (2) at least one device user interface functionality, the (1) at least one device user interface element and/or (2) at least one device user interface functionality based, at least in part, on data and/or information resulting from one or any combination of methods or processes disclosed in this application as relevant to any embodiment of the invention, and/or at least one signal resulting from one or any combination of methods (or processes) disclosed in this application as relevant to any embodiment of the invention.

[0011] For various example embodiments of the invention, the following is also applicable: a method comprising creating and/or modifying (1) at least one device user interface element and/or (2) at least one device user interface functionality, the (1) at least one device user interface element and/or (2) at least one device user interface functionality based at least in part on data and/or information resulting from one or any combination of methods (or processes) disclosed in this application as relevant to any embodiment of the invention, and/or at least one signal resulting from one or any combination of methods (or processes) disclosed in this application as relevant to any embodiment of the invention.

[0012] In various example embodiments, the methods (or processes) can be accomplished on the service provider side or on the mobile device side or in any shared way between service provider and mobile device with actions being performed on both sides.

[0013] For various example embodiments, the following is applicable: An apparatus comprising means for performing a method of the claims.

[0014] Still other aspects, features, and advantages of the invention are readily apparent from the following detailed description, simply by illustrating a number of particular embodiments and implementations, including the best mode contemplated for carrying out the invention. The invention is also capable of other and different embodiments, and its several details can be modified in various obvious respects, all without departing from the spirit and scope of the invention. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.

BRIEF DESCRIPTION OF THE DRAWINGS



[0015] The embodiments of the invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings:

FIG. 1 is a diagram of a system capable of determining a ground control point from image data, according to one embodiment;

FIG. 2 is a diagram of components of a machine learning system capable of determining a ground control point from image data, according to one embodiment;

FIG. 3 is a flowchart of a process for determining a ground control point from image data, according to one embodiment;

FIG. 4 is a diagram illustrating example intersection features, according to one embodiment;

FIGs. 5A-5C are diagrams illustrating example imagery of intersection features, according to one embodiment;

FIGs. 6A and 6B are diagrams illustrating multiple images of the same intersection feature, according to one embodiment;

FIG. 7 is a flowchart of a process for generating prediction location data for a ground control point using a trained machine learning model, according to one embodiment;

FIG. 8 is a diagram illustrating a workflow for using ground control points, according to one embodiment;

FIG. 9 is a diagram of a geographic database, according to one embodiment;

FIG. 10 is a diagram of hardware that can be used to implement an embodiment;

FIG. 11 is a diagram of a chip set that can be used to implement an embodiment; and

FIG. 12 is a diagram of a mobile terminal (e.g., handset) that can be used to implement an embodiment.


DESCRIPTION OF SOME EMBODIMENTS



[0016] Examples of a method, apparatus, and computer program for determining a ground control point from image data using machine learning are disclosed. In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the invention. It is apparent, however, to one skilled in the art that the embodiments of the invention may be practiced without these specific details or with an equivalent arrangement. In other instances, well-known structures and devices are shown in block diagram form in order to avoid unnecessarily obscuring the embodiments of the invention.

[0017] FIG. 1 is a diagram of a system capable of determining a ground control point from image data, according to one embodiment. As indicated above, many location-based services and applications rely on accurate map data. For example, automated driving is quickly becoming a reality following advances in machine learning, computer vision, and compute power. The ability to perceive the world with an accurate semantic understanding enables vehicles (e.g., an autonomous vehicle 101) to obey driving rules and avoid collisions. As these perceptual abilities have improved, so too has the need for highly accurate and up-to-date maps. Path planning, for instance, requires knowledge of what to expect beyond a vehicle 101's perceptual horizon, and driving in complicated urban environments with many occluding objects requires a knowledge of what cannot be detected by onboard sensors.

[0018] In response, map service providers (e.g., operating a mapping platform 103) are creating the most accurate and up-to-date high-resolution map for automated driving (e.g., a geographic database 105). To facilitate and/or monitor the accuracy of digital map data stored in the geographic database 105, map service providers can designate ground control points. In one embodiment, ground control points are defined as identifiable points on the Earth's surface that have precise location (e.g., in the form of <Latitude, Longitude, Elevation>) associated with them. These points play a vital role in being able to measure the quality and correction of different data sources.

[0019] In other embodiments, ground control points find additional applications in camera pose refinement of satellite, aerial and ground imagery, and hence provide for increased position fidelity for location data determined from these data sources. In turn, any derived products like building polygons, map objects made from these data sources inherit the accuracy. In addition, ground control points can also serve the localization of the automated car use case where they can be geocoded localization objects that car can measure its position with respect to.

[0020] Traditionally, ground control points are collected by ground surveyors who go out in the field and use instruments like a theodolite, measuring tape, three-dimensional (3D) scanner, satellite-based location sensors (e.g., GPS/GNSS), level and rod, etc. to measure the locations of ground control points with respect to the locations of distinguishable landmarks on the Earth (e.g. parts of signs, barriers, buildings, road paint, etc.). Collecting each ground control point using traditional manual means requires a substantial amount of infrastructure and manual resources. The problems become even more pronounced if the ground control points need to be measured on the road (e.g., for map making use cases) since special access permissions need to be obtained from the government or other responsible authorities. Because of the infrastructure and resource burden, the process of obtaining ground control points using traditional means is not scalable if they need to be used in map making and evaluation process.

[0021] To complicate the process further, ground control points are valid for unpredictable periods of time. For example, a previously measure ground control point can become invalid or obsolete if the feature or object (e.g., a building, lane marking, etc.) on which the ground control point is based changes, for instance, due to construction, paint deterioration, and/or other changes to the environment. Other changes, for instance, can include to shifts in tectonic plates or other geological movements that shift the location of ground control points by a couple of centimeters or more per year. For high definition map use (e.g., with centimeter level accuracy), those micro changes in ground control points can have an effect on the accuracy of digital maps. Accordingly, map service providers face significant technical challenges to determining ground control points that can scale (e.g., with increased map coverage) given limited available resources and that can be updated at a frequency sufficient to reduce the probability of a ground control point becoming invalid or obsolete below a target threshold.

[0022] To address these technical challenges and problems, the system 100 of FIG. 1 introduces a capability to identify ground control points that are learnable from top down imagery by machine learning systems (e.g., using the machine learning system 107 of the mapping platform 103 in combination with the computer vision system 109). In one embodiment, the system 100 can mark or label these learnable ground control points (e.g., an intersection of ground paint lines) in a large of set of training images to train the machine learning system 107. The trained machine learning system 107 can then be used to automatically find the locations of all such ground control points in new images. By way of example, such imagery or image data can be obtained from different sources such as but not limited to satellites, airplanes, drones, and/or other aerial vehicles. In one embodiment, the system 100 selects the map features for use by the machine learning system 107 (e.g., for training, evaluation, and/or prediction) based on determining whether the map features are "learnable" by the machine learning system 107. Learnable, for instance, refers to whether the ground control point/map feature can be used to train the machine learning system 107 to make predictions of a corresponding location of the ground control point as depicted in image data (e.g., top-down imagery).

[0023] In one embodiment, to determine whether a ground control point/map feature is learnable, the system 100 can determine whether the map feature satisfies any combination of designated properties such as but not limited to:
  • The ground control point/map feature should have a consistent definition so that the machine learning system 107 (e.g., machine learning models) can learn it;
  • The ground control point/map feature should be uniquely identifiable;
  • The ground control point/map feature should be spatially sparse so the correspondence task can be simplified; and
  • The definition of the ground control point/map feature should be generalizable enough be applicable in different parts or regions of the world.


[0024] After selecting learnable map features/ground control points according to the criteria above, the system 100 can label image data (e.g., top-down imagery) depicting examples of the selected ground control points to generate a set of training images. Labeling, for instance, includes identifying (e.g., using a human labeler) the pixel locations within each image corresponding to the learnable ground control points present in the image. The system 100 can then use the labeled training images to train the machine learning system 107 to predict all such ground control points in the input image data (e.g., by predicting the pixel locations or data indicating the pixel locations of the ground control points in the input images).

[0025] In one embodiment, the system 100 uses the trained machine learning system 107 to automatically label the ground control points in new imagery depicting areas to be mapped or analyzed. Then, the 3D position of the ground control point/map feature can be estimated from multiple top-down views, with corresponding ground control points labeled in two or more images. For example, the system 100 can determine pixel correspondences between the ground control points labeled in each of the images. The 3D position of the ground control points can then be determined via a triangulation process from the pixel correspondences of each ground control point in combination with a camera model or camera pose (e.g., camera position, pointing direction, etc.) of the camera system used to capture the imager. In addition, different image sources (e.g., satellites, airplanes, drones, etc.) provide imagery with different qualities and resolutions and hence the uncertainty/error associated with the generated ground control points from them will be different and can also be computed. The embodiments for determining ground control points from image data using machine learning are described in more detail below.

[0026] As shown in FIG. 1, in one embodiment, the mapping platform 103 includes the machine learning system 107 and computer vision system 109 for determining ground control points from image data. FIG. 2 is a diagram of components of the machine learning system 107 that includes one or more components for determining ground control points from image data according to the various embodiments described herein. It is contemplated that the functions of these components may be combined or performed by other components of equivalent functionality. In this embodiment, the machine learning system 107 includes a feature module 201, , a training module 203, a prediction module 205, and a triangulation module 207. The above presented modules and components of the machine learning system 107 can be implemented in hardware, firmware, software, or a combination thereof. Though depicted as a component of the mapping platform 103 in FIG. 1, it is contemplated that the machine learning system 107 may be a separate entity or may be implemented as a module of any other component of the system 100 (e.g., a component of the computer vision system 109, services platform 111, services 113a-113n (also collectively referred to as services 113), vehicle 101, a user equipment (UE) 115, application 117 executing on the UE 115, etc.). In another embodiment, one or more of the modules 201-207 may be implemented as a cloud based service, local service, native application, or combination thereof. The functions of the machine learning system 107 and the modules 201-207 are discussed with respect to FIGs. 3-8 below.

[0027] FIG. 3 is a flowchart of a process 300 for determining a ground control point from image data, according to one embodiment. More specifically, the embodiments of the process 300 can be used to train a machine learning model of the system 100 to predict the locations and/or other characteristics of ground control points depicted in image data. In various embodiments, the machine learning system 107 and/or any of the modules 201-207 of the machine learning system 107 may perform one or more portions of the process 300 and may be implemented in, for instance, a chip set including a processor and a memory as shown in FIG. 11. As such, the machine learning system 107 and/or the modules 201-207 can provide means for accomplishing various parts of the process 300, as well as means for accomplishing embodiments of other processes described herein in conjunction with other components of the system 100. Although the process 300 is illustrated and described as a sequence of steps, its contemplated that various embodiments of the process 300 may be performed in any order or combination and need not include all of the illustrated steps.

[0028] In step 301, the feature module 201 selects or receives an input (e.g., from a user or system administrator) for selecting features that are to be used for designating ground control points. Features, for instance, refer to any physical feature identifiable on the ground as possible candidates. However, not all of the available candidate features are suitable for machine learning (i.e., "learnable") as discussed above. Accordingly, in one embodiment, the feature selection module 201 can evaluate or receive input specifying an evaluation of the candidate features for selection as ground control points.

[0029] As previously discussed, in one embodiment, the feature module 201 can use one or more designated properties for determining whether a candidate feature is learnable and therefore can be selected as a ground control point for machine learning. For example, the designated properties can include but are not limited to: (1) having a consistent definition, (2) being uniquely identifiable, (3) having spatial sparsity, and/or (4) being generalizable across different geographic regions. In one embodiment, the feature selection process can be performed as part of an initial set up phase. By way of example, the category of curvilinear geometry intersections includes physical features (e.g., intersection features) which generally meet the above criteria to be candidates as machine learnable ground control points. Curvilinear geometry intersection features are features defined by lines, markings, structures, etc. that are found at roadway intersections. The features can also include any geometric arrangement of the features (e.g., line intersections, angles, boundaries, etc.).

[0030] In one embodiment, the intersection features are those features or geometric arrangements of the features that are visible in top-down imagery. It is noted that intersection features or curvilinear geometry features are provided as examples of physical features that can meet the learnable criteria described above. Although the various embodiments are discussed with respect to intersection feature, it is contemplated that any feature that meets the above criteria can be used as learnable ground control points according to the embodiments described herein. As such, the computer vision system 109 can more easily identify these features using object recognition on the corresponding images. Top-down imagery refers to images or image data that are captured from an overhead or aerial perspective so that the camera is pointed down towards the intersection or ground level from an overhead height. The axis of the pointing direction of the camera can vary from a direct overhead (e.g., perpendicular angle) or to an oblique angle from either side. In one embodiment, the camera pose or position data can be provided with the imagery and then refined to greater accuracy using ground control points. Other camera attributes (e.g., focal length, camera type, etc.) and/or environmental attributes (e.g., weather, time of day, etc.) can be provided with the imagery.

[0031] FIG. 4 is a diagram illustrating example intersection features that can be found at a typical intersection, according to one embodiment. The example of FIG. 4 illustrates a schematic drawing of a typical intersection 400 at which intersection features created by various geometries of the lane lines, crosswalks, bus stops, and/or any other identifiable object or marking found at the intersection 400. An intersection refers, for instance, to a geographic area at which two or more road segments intersect, converge, and/or diverge. As shown, intersection features in the category of curvilinear geometry include but are not limited to:
  1. (1) Points 401a-401c at which a lane boundary (e.g., lane line or marking) meets a crosswalk;
  2. (2) Points 403a and 403 that correspond to the corners of road markings indicating a bus stop;
  3. (3) Points 405a-405d that correspond to the corners of a crosswalk;
  4. (4) Points 407a and 407b that are the top of gore points touching a crosswalk or limit lines (e.g., lines designating the limit or boundaries of other features such as lanes); and
  5. (5) Point 409 at which a limit line meets a lane boundary.


[0032] The intersection features identified above typically meet the criteria or properties for being classified as learnable according to the embodiments described herein. For example, the property of having a "consistent definition" (e.g., see designated property item (1) above) indicates whether a feature description can be consistently applied and learned by the machine learning system 107. This consistency can depend on whether the machine learning system 107 can identify the feature or ground control point based on its definition with accuracy above a threshold value. For example, a map or intersection feature can be defined using a descriptive property such as "a lane boundary with a crosswalk". This means that the map feature/ground control point would correspond to the location or point at which a lane line meets a cross walk line. In one embodiment, the consistency of the definition of the ground control point can be measured by determining whether the feature module 201 can correctly identify the feature in ground truth images above a threshold accuracy level (e.g., more than 80% correct). Map or intersection features that meet this threshold can then be classified as having a consistent definition that satisfies the property or criterion.

[0033] In one embodiment, the intersection feature/ground control point is selected so that the intersection feature is uniquely identifiable from among other intersection features from the category of curvilinear geometry intersection features (e.g., see designated property item (2) above). In other words, a single feature should only be classified under one feature definition category. For example, if a feature that is classified as "a lane boundary with a crosswalk" should also not satisfy the definition for being a "gore point" or vice versa. In one embodiment, if a feature is not uniquely identifiable, the feature module 201 can classify the feature as not learnable.

[0034] In another embodiment, the intersection feature is selected based on determining that the intersection feature has a spatial sparsity that meets a sparsity criterion (e.g., see designated property item (3) above). Under this property, the feature module 201 can determine the number of same intersection feature occurring in a known geographic to calculate the spatial sparsity of the feature. The feature module 201 can then classify the feature as learnable if the sparsity (e.g., number features per area) is below the sparsity criterion or threshold. In contrast, features that repeat often within an designated area (i.e., not sparse or appear in numbers greater than the sparsity threshold) are not well suited as ground control points because they can be more difficult to uniquely identify and match as against known ground control points. For example, features such as dashes of a lane line, stripes in a crosswalk, multiple line paint intersections in restricted zones, zebra stripes, etc. that repeat often over small distances can be poor ground control point candidates.

[0035] In yet another embodiment, the feature module 201 can determine whether the intersection feature is applicable to a plurality of different geographic regions (e.g., see designated property item (1) above). In other words, features are learnable if they are likely to appear across various regions of the world. If a selected feature or ground control point is specific only to a particular area, then the resulting trained machine learning model would be applicable only to the particular area. To provide a more general machine learning system 107, the feature module 201 can select only those features that occur in all or a majority (e.g., greater than a threshold percentage) of the areas or regions that are of interest.

[0036] FIGs. 5A-5C illustrate example imagery of some of the intersection features illustrated in FIG. 4, according to one embodiment. For example, FIG. 5A illustrates top-down imagery 500 that depicts ground control points 501a-501e at which a lane boundary meets a crosswalk. FIG. 5B illustrates to-down imagery 520 that depicts ground control points 521a and 521b that are crosswalk corners. FIG. 5C illustrates top-down imagery 540 that depicts ground control points 541a and 541b at which a limit line meets a lane line. Each of the ground control points illustrated in FIGs. 5A-5C are an intersection features from a category of curvilinear geometry intersection features that are a visible feature of a roadway intersection (e.g., visible from a top-down imagery perspective).

[0037] After selecting the features that are to be designated as ground control points, the feature module 201 can label and/or retrieve a plurality of ground truth images depicting the intersection feature (step 303). In one embodiment, the plurality of ground truth images is labeled with known pixel location data of the intersection as respectively depicted in each of the plurality of ground truth images. The known pixel location data indicate which pixel(s) of a ground truth image correspond to ground control points that are present in the image. As previously described, the known pixel location data can be used to determined pixel correspondences between multiple images to determine real world three-dimensional locations of the ground control point (e.g., intersection feature) comprising a latitude, longitude, and elevation. The ground truth images can also include multiple images of the same ground control point or learnable feature (e.g., captured at different times, from different sources, etc.).

[0038] In one embodiment, to determine or label pixel location data, the feature module 201 can process the images using image recognition or equivalent to identify the pixels of each image corresponding to the selected intersection features/ground control points. In other words, following the identification of candidate feature points in several top-down images, corresponding image pixel locations are identified. In one embodiment, for each real-world feature (e.g., line intersection), the corresponding pixel coordinates in two or more images are recorded, creating a pixel correspondence of the form {(u1, v1), (u2, v2), ...} or equivalent. Here, u and v are pixel locations of the same physical object or feature depicted in the images (e.g., pixel locations along the x and y axis respectively of a pixel grid comprising the image), and the subscript indicates in which image the feature is labeled.

[0039] FIGs. 6A and 6B are diagrams illustrating multiple images of the same intersection feature, according to one embodiment. In this example, FIG. 6A illustrates a first top-down image 601 that depicts an intersection with a crosswalk corner 603 in a main view 605 and a zoomed view 607, and FIG. 6B illustrates a second top-down image 621 of the same intersection with the crosswalk corner 603 in the main view 605 and the zoomed view 607 captured at a different time. FIGs. 6A and 6B illustrate an example of a crosswalk corner pixel correspondence (e.g., for crosswalk corner 603) between two satellite image patches. In one embodiment, the pixel correspondences, together with image metadata such as but not limited to camera position, orientation, focal length, etc. can be used to estimate the 3D position of the intersection feature/ground control point (e.g., the crosswalk corner 603).

[0040] Once the definition of such ground control points is done, a large set of annotated or human-created observations (e.g., ground truth images of intersections features) could be obtained. For example, to generate the ground truth data, correspondences among detected points can be determined by human labelers (perhaps with visual aids to orient and co-register multiple images) or through automated means (brute-force matching, approximate nearest neighbors, supervised deep neural networks, etc.). The output of this process is a set of ground truth images labeled with learnable map features/ground control points.

[0041] In step 305, the training module 203 can present this ground truth image data to a machine learning model of the machine learning system 107 during training using, for instance, supervised deep convolutional networks or equivalent. In other words, the training module 305 trains a machine learning model using the plurality of ground truth images to identify learnable ground control points depicted in an input images. Generally, a machine learning model (e.g., a neural network, set of equations, rules, decision trees, etc.) is trained to manipulate an input feature set to make a prediction about the feature set or the phenomenon/observation that the feature set represents. In one embodiment, the training features for the machine learning model include the determined pixel correspondence or pixel location of the selected map features/ground control points in the ground truth images.

[0042] In one embodiment, because the ground truth images can originate from any number of sources, the resolution, quality, etc. of each image can vary. For example, the resolution of top imagery of different satellites or other aerial sources can vary depending on the kind of camera sensors used. These different sensors then produce images with different resolutions. This variance, in turn, can lead to uncertainty or error. Accordingly, the machine learning model can be further trained to calculate an uncertainty associated with the predicted location based on a characteristic of said each of the plurality of images, a respective source of said each of the plurality of images, or a combination.

[0043] In one embodiment, the training module 203 can incorporate a supervised learning model (e.g., a logistic regression model, RandomForest model, and/or any equivalent model) to train a machine learning model using the ground truth image data together with the labeled ground control points. For example, during training, the training module 203 uses a learner module that feeds images and derived feature sets (e.g., pixel correspondences, image attributes, etc.) into the machine learning model to compute a predicted feature set (e.g., predicted ground control points presented in input images and/or other characteristics of ground control points) using an initial set of model parameters.

[0044] The learner module then compares the predicted feature set to the ground truth data (e.g., images labeled with known ground control point pixel locations and/or attributes). For example, the learner module computes a loss function representing, for instance, an accuracy of the predictions for the initial set of model parameters. In one embodiment, the training module 203 computes a loss function for the training of the machine learning module based on the ground truth images. The learner module of the training module 203 then incrementally adjusts the model parameters until the model minimizes the loss function (e.g., achieves a maximum accuracy with respect to the manually marked labels). In other words, a "trained" feature prediction model is a classifier with model parameters adjusted to make accurate predictions with respect to the ground truth data.

[0045] FIG. 7 is a flowchart of a process for predicting ground control point using in images using a trained machine learning model, according to one embodiment. In various embodiments, the machine learning system 107 and/or any of the modules 201-207 of the machine learning system 107 may perform one or more portions of the process 300 and may be implemented in, for instance, a chip set including a processor and a memory as shown in FIG. 11. As such, the machine learning system 107 and/or the modules 201-207 can provide means for accomplishing various parts of the process 300, as well as means for accomplishing embodiments of other processes described herein in conjunction with other components of the system 100. Although the process 300 is illustrated and described as a sequence of steps, its contemplated that various embodiments of the process 300 may be performed in any order or combination and need not include all of the illustrated steps.

[0046] In one embodiment, the process 700 can be used to classify or identify ground control points in image data (e.g., top-down imagery) after a machine learning model of the machine learning system 107 is trained (e.g., according to the embodiments of the process 300 above).

[0047] In step 701, the prediction module 205 retrieves or is provided with one or more input images that depict potential ground control points that are learnable features (e.g., intersection features) according to the embodiments described above. In one embodiment, the input image data include top-down imagery (e.g., captured by cameras mounted on satellites, airplanes, drones, and/or other aerial vehicles). The prediction module 205 can process the input image data to identify ground control points (e.g., learnable intersection features). The processing can include, for instance, using the computer vision system 109 or equivalent to recognize pixels corresponding to the features/ground control points of interest. The recognition process can identify the pixel locations in each image corresponding to the features/ground control points. If the input image data include multiple images (e.g., two or more images) of the same feature or ground control point, the prediction module 205 can create pixel correspondences of the feature across the multiple images according to the embodiments described with respect to FIG. 3 above.

[0048] In step 703, the prediction module 205 creates an input vector, input matrix, or equivalent comprising the pixel locations/pixel correspondences extracted above along with any other features/attributes used to train the machine learning model. By way of example, these other features/attributes can include but is not limited to image itself, derived attributes of the images (e.g., resolution, exposure, camera position, focal length, camera type, etc.), the corresponding data sources (e.g., satellite, airplane, drone, etc.), contextual data at the time of image capture (e.g., time, weather, etc.), and/or the like. In one embodiment, creating the input vector includes converting the extracted features to a common format, normalizing values, removing outliers, and/or any other known pre-processing step to reduce data anomalies in the input data.

[0049] In step 705, the prediction module 205 processes the input vector using the trained machine learning model to output a predicted pixel location of the ground control point/feature that is present in the input image. In one embodiment, the predicted ground control point data can also include a predicted uncertainty or error associated with the predicted ground control point (e.g., error in predicted pixel locations of the ground control point).

[0050] In one embodiment, the ground control points once predicted and obtained using machine learning could serve multiple purposes including but not limited to evaluation of map data quality, absolute constraints for the positional correction of data sources, serve as localization objects, etc. For example, in one embodiment, the triangulation module 207 can determine pixel correspondences between the same ground control point automatically labeled in multiple images. The triangulation module 207 can then use a triangulation process to estimate a real-world 3D location of the ground control points from the pixel correspondences and camera model or camera pose information. The triangulation process can use the known camera location, pointing direction, and/or other camera attributes (e.g., focal length, etc.) indicated in the camera model or pose information to triangulate the real-world location of the ground control point from the known camera locations of different images in which the ground control is labeled (e.g., labeled according to the embodiments of the automated machine learning process described above).

[0051] FIG. 8 is a diagram illustrating a workflow for using ground control points, according to one embodiment. As shown in the process 800 of FIG. 8, a set of input images 801 (e.g., top-down imagery) is obtained, and then processed using the trained machined learning model 803 (e.g., of the machine learning system 107) to generate ground control points (e.g., ground control point data 805). This ground control point data 805 can then be provided to the mapping platform 103 and/or any other component of the system 100 (e.g., services platform 111, services 113, vehicle 101, etc.) for the purposes listed above or other purposes that rely on ground control points. Because capturing images is generally a less resource intensive process than deploying ground surveyors to manually determine ground control points, the system 100 can capture a series of input images 801 to cover a wider geographical area more frequently for automated processing by the machine learning system 107. As a result, the embodiments described herein for determining ground control points from image data using machine learning can advantageously provide for scalable, relatively inexpensive, high accuracy ground control point data 805 that is also easy to keep up-to-date (e.g., thereby reducing invalid or obsolete ground control points).

[0052] In one embodiment, the embodiments of the machine learning system 107 can be used to enable a variety of sophisticated services and applications. For example, autonomous driving has quickly become an area of intense interest where machine learning in combination with computer vision systems can be used. One application of vision techniques in autonomous driving is localization of the vehicle 101 with respect to ground control points (e.g., reference locations with highly accurate known locations). In one embodiment, the system 100 (e.g., the mapping platform 103) can generate ground control points according to the embodiments as described herein. These ground control points can then be used as reference markers by vehicles 101 to localize themselves.

[0053] Traditionally, most vehicle navigation systems have accomplished this localization using GPS, which generally provides a real-time location with a 95% confidence interval of 7.8 meters. However, in complicated urban environments, reflection of GPS signals can further increase this error, such that one's location may be off by as much as 30 meters. Given that the width of many lanes is 3-4 meters, this accuracy is not sufficient to properly localize a vehicle 101 (e.g., an autonomous vehicle) so that it can make safe route planning decisions. Other sensors, such as inertial measurement units (IMUs) can increase the accuracy of localization by taking into account vehicle movement, but these sensors tend to drift and still do not provide sufficient accuracy for localization.

[0054] In general, a localization accuracy of around 10 cm is needed for safe driving (e.g., autonomous driving) in many areas. One way to achieve this level of accuracy is to use visual odometry, in which features (e.g., ground control points) are detected from imagery using feature prediction models (i.e., a machine learning classifier). These features can then be matched to a database of ground control points to determine one's location. By way of example, traditional feature-based localization that both detect features and localize against them generally rely on low-level features. However, low-level features typically used in these algorithms (e.g., Scale-Invariant Feature Transform (SIFT) or Oriented FAST and rotated BRIEF (ORB)) tend to be brittle and not persist in different environmental and lighting conditions. As a result, they often cannot be used to localize a vehicle on different days in different weather conditions. Aside from reproducibility, the ability to detect and store higher level features of different types (e.g., ground control points based on intersection features such as lane markings, lane lines, etc.) can provide better and more accurate localization.

[0055] A vehicle 101, for instance, can use computer vision to identify a known ground control point (e.g., a crosswalk corner), and then estimate its distance to the ground control point. Because the location of the ground control point is known with high accuracy, the vehicle 101 can compute its distance to the ground control point to use as a distance offset to the known location to localize itself with a corresponding high degree of accuracy. Understanding one's location on a map enables planning of a route, both on fine and coarse scales. On a coarse scale, navigation maps allow vehicles 101 to know what roads to use to reach a particular destination. However, on a finer scale, maps allow vehicles 101 to know what lanes to be in and when to make lane changes. Knowing this information is important for planning an efficient and safe route, for in complicated driving situations maneuvers need to be executed in a timely fashion, and sometimes before they are visually obvious. In addition, localization with respect to ground control points enables the incorporation of other real-time information into route planning. Such information could include traffic, areas with unsafe driving conditions (ice, fog, potholes, e.g.), and temporary road changes like construction.

[0056] Returning to FIG. 1, as shown, the system 100 includes the mapping platform 103 which incorporates the machine learning system 107 for determining ground control points from image data according the various embodiments described herein. In addition, the mapping platform 103 can include the computer vision system 109 configured to use machine learning to detect objects or features (e.g., intersection features) depicted in images that can be used as ground control points. For example, with respect to autonomous, navigation, mapping, and/or other similar applications, the computer vision system 109 can detect ground control points in input images and generate ground control point data (e.g., location data) and associated prediction confidence values/uncertainties, according to the various embodiments described herein. In one embodiment, the machine learning system 107 includes a neural network or other machine learning system to make predictions from machine learning models. For example, when the input to the machine learning model are images used for visual odometry, the features of interest can include ground control points detected in the images to support localization of, e.g., a vehicle 101 or other similar applications within the sensed environment. In one embodiment, the neural network of the machine learning system 107 is a traditional convolutional neural network which consists of multiple layers of collections of one or more neurons (which are configured to process a portion of an input image. In one embodiment, the receptive fields of these collections of neurons (e.g., a receptive layer) can be configured to correspond to the area of an input image delineated by a respective a grid cell generated as described above.

[0057] In one embodiment, the machine learning system 107 and/or the computer vision system 109 also have connectivity or access to a geographic database 105 which stores the learned ground control points generated according to the embodiments described herein. In one embodiment, the geographic database 105 includes representations of mapped ground control points and related geographic features to facilitate visual odometry to increase localization accuracy. In one embodiment, the machine learning system 107 and/or computer vision system 109 have connectivity over a communication network 119 to the services platform 111 that provides one or more services 113. By way of example, the services 113 may be third party services and include mapping services, navigation services, travel planning services, notification services, social networking services, content (e.g., audio, video, images, etc.) provisioning services, application services, storage services, contextual information determination services, location based services, information based services (e.g., weather, news, etc.), etc. In one embodiment, the services 113 uses the output of the machine learning system 107 and/or of the computer vision system 109 (e.g., ground control point data) to localize the vehicle 101 or UE 115 (e.g., a portable navigation device, smartphone, portable computer, tablet, etc.) to provide services 113 such as navigation, mapping, other location-based services, etc.

[0058] In one embodiment, the machine learning system 107 and/or computer vision system 109 may be a platform with multiple interconnected components. The machine learning system 107 and/or computer vision system 109 may include multiple servers, intelligent networking devices, computing devices, components and corresponding software for providing parametric representations of lane lines. In addition, it is noted that the machine learning system 107 and/or computer vision system 109 may be a separate entity of the system 100, a part of the one or more services 113, a part of the services platform 111, or included within the UE 115 and/or vehicle 101.

[0059] In one embodiment, content providers 121a-121m (collectively referred to as content providers 121) may provide content or data (e.g., including geographic data, parametric representations of mapped features, etc.) to the geographic database 105, the machine learning system 107, the computer vision system 109, the services platform 111, the services 113, the UE 115, the vehicle 101, and/or an application 117 executing on the UE 115. The content provided may be any type of content, such as map content, textual content, audio content, video content, image content, etc. In one embodiment, the content providers 121 may provide content that may aid in the detecting and classifying of lane lines and/or other features in image data and estimating the quality of the detected features. In one embodiment, the content providers 121 may also store content associated with the geographic database 105, machine learning system 107, computer vision system 109, services platform 111, services 113, UE 115, and/or vehicle 101. In another embodiment, the content providers 121 may manage access to a central repository of data, and offer a consistent, standard interface to data, such as a repository of the geographic database 105.

[0060] In one embodiment, the UE 115 and/or vehicle 101 may execute a software application 117 to capture image data or other observation data for determining ground control points or using ground control points according the embodiments described herein. By way of example, the application 117 may also be any type of application that is executable on the UE 115 and/or vehicle 101, such as autonomous driving applications, mapping applications, location-based service applications, navigation applications, content provisioning services, camera/imaging application, media player applications, social networking applications, calendar applications, and the like. In one embodiment, the application 117 may act as a client for the machine learning system 107 and/or computer vision system 109 and perform one or more functions associated with determining ground control points from image data alone or in combination with the machine learning system 107.

[0061] By way of example, the UE 115 is any type of embedded system, mobile terminal, fixed terminal, or portable terminal including a built-in navigation system, a personal navigation device, mobile handset, station, unit, device, multimedia computer, multimedia tablet, Internet node, communicator, desktop computer, laptop computer, notebook computer, netbook computer, tablet computer, personal communication system (PCS) device, personal digital assistants (PDAs), audio/video player, digital camera/camcorder, positioning device, fitness device, television receiver, radio broadcast receiver, electronic book device, game device, or any combination thereof, including the accessories and peripherals of these devices, or any combination thereof. It is also contemplated that the UE 115 can support any type of interface to the user (such as "wearable" circuitry, etc.). In one embodiment, the UE 115 may be associated with the vehicle 101 or be a component part of the vehicle 101.

[0062] In one embodiment, the UE 115 and/or vehicle 101 are configured with various sensors for generating or collecting environmental image data (e.g., for processing by the machine learning system 107 and/or computer vision system 109), related geographic data, etc. In one embodiment, the sensed data represent sensor data associated with a geographic location or coordinates at which the sensor data was collected. By way of example, the sensors may include a global positioning sensor for gathering location data (e.g., GPS), a network detection sensor for detecting wireless signals or receivers for different short-range communications (e.g., Bluetooth, Wi-Fi, Li-Fi, near field communication (NFC) etc.), temporal information sensors, a camera/imaging sensor for gathering image data (e.g., the camera sensors may automatically capture ground control point imagery, etc. for analysis), an audio recorder for gathering audio data, velocity sensors mounted on steering wheels of the vehicles, switch sensors for determining whether one or more vehicle switches are engaged, and the like.

[0063] Other examples of sensors of the UE 115 and/or vehicle 101 may include light sensors, orientation sensors augmented with height sensors and acceleration sensor (e.g., an accelerometer can measure acceleration and can be used to determine orientation of the vehicle), tilt sensors to detect the degree of incline or decline of the vehicle along a path of travel, moisture sensors, pressure sensors, etc. In a further example embodiment, sensors about the perimeter of the UE 115 and/or vehicle 101 may detect the relative distance of the vehicle from a lane or roadway, the presence of other vehicles, pedestrians, traffic lights, potholes and any other objects, or a combination thereof. In one scenario, the sensors may detect weather data, traffic information, or a combination thereof. In one embodiment, the UE 115 and/or vehicle 101 may include GPS or other satellite-based receivers to obtain geographic coordinates from satellites 123 for determining current location and time. Further, the location can be determined by visual odometry, triangulation systems such as A-GPS, Cell of Origin, or other location extrapolation technologies. In yet another embodiment, the sensors can determine the status of various control elements of the car, such as activation of wipers, use of a brake pedal, use of an acceleration pedal, angle of the steering wheel, activation of hazard lights, activation of head lights, etc.

[0064] In one embodiment, the communication network 119 of system 100 includes one or more networks such as a data network, a wireless network, a telephony network, or any combination thereof. It is contemplated that the data network may be any local area network (LAN), metropolitan area network (MAN), wide area network (WAN), a public data network (e.g., the Internet), short range wireless network, or any other suitable packet-switched network, such as a commercially owned, proprietary packet-switched network, e.g., a proprietary cable or fiber-optic network, and the like, or any combination thereof. In addition, the wireless network may be, for example, a cellular network and may employ various technologies including enhanced data rates for global evolution (EDGE), general packet radio service (GPRS), global system for mobile communications (GSM), Internet protocol multimedia subsystem (IMS), universal mobile telecommunications system (UMTS), etc., as well as any other suitable wireless medium, e.g., worldwide interoperability for microwave access (WiMAX), Long Term Evolution (LTE) networks, code division multiple access (CDMA), wideband code division multiple access (WCDMA), wireless fidelity (Wi-Fi), wireless LAN (WLAN), Bluetooth®, Internet Protocol (IP) data casting, satellite, mobile ad-hoc network (MANET), and the like, or any combination thereof.

[0065] By way of example, the machine learning system 107, computer vision system 109, services platform 111, services 113, UE 115, vehicle 101, and/or content providers 121 communicate with each other and other components of the system 100 using well known, new or still developing protocols. In this context, a protocol includes a set of rules defining how the network nodes within the communication network 119 interact with each other based on information sent over the communication links. The protocols are effective at different layers of operation within each node, from generating and receiving physical signals of various types, to selecting a link for transferring those signals, to the format of information indicated by those signals, to identifying which software application executing on a computer system sends or receives the information. The conceptually different layers of protocols for exchanging information over a network are described in the Open Systems Interconnection (OSI) Reference Model.

[0066] Communications between the network nodes are typically effected by exchanging discrete packets of data. Each packet typically comprises (1) header information associated with a particular protocol, and (2) payload information that follows the header information and contains information that may be processed independently of that particular protocol. In some protocols, the packet includes (3) trailer information following the payload and indicating the end of the payload information. The header includes information such as the source of the packet, its destination, the length of the payload, and other properties used by the protocol. Often, the data in the payload for the particular protocol includes a header and payload for a different protocol associated with a different, higher layer of the OSI Reference Model. The header for a particular protocol typically indicates a type for the next protocol contained in its payload. The higher layer protocol is said to be encapsulated in the lower layer protocol. The headers included in a packet traversing multiple heterogeneous networks, such as the Internet, typically include a physical (layer 1) header, a data-link (layer 2) header, an internetwork (layer 3) header and a transport (layer 4) header, and various application (layer 5, layer 6 and layer 7) headers as defined by the OSI Reference Model.

[0067] FIG. 9 is a diagram of a geographic database, according to one embodiment. In one embodiment, the geographic database 105 includes geographic data 901 used for (or configured to be compiled to be used for) mapping and/or navigation-related services, such as for video odometry based on the mapped features (e.g., lane lines, road markings, signs, etc.). In one embodiment, the geographic database 105 includes high resolution or high definition (HD) mapping data that provide centimeter-level or better accuracy of map features. For example, the geographic database 105 can be based on Light Detection and Ranging (LiDAR) or equivalent technology to collect billions of 3D points and model road surfaces and other map features down to the number lanes and their widths. In one embodiment, the HD mapping data (e.g., HD data records 911) capture and store details such as the slope and curvature of the road, lane markings, roadside objects such as sign posts, including what the signage denotes. By way of example, the HD mapping data enable highly automated vehicles to precisely localize themselves on the road.

[0068] In one embodiment, geographic features (e.g., two-dimensional or three-dimensional features) are represented using polygons (e.g., two-dimensional features) or polygon extrusions (e.g., three-dimensional features). For example, the edges of the polygons correspond to the boundaries or edges of the respective geographic feature. In the case of a building, a two-dimensional polygon can be used to represent a footprint of the building, and a three-dimensional polygon extrusion can be used to represent the three-dimensional surfaces of the building. It is contemplated that although various embodiments are discussed with respect to two-dimensional polygons, it is contemplated that the embodiments are also applicable to three-dimensional polygon extrusions. Accordingly, the terms polygons and polygon extrusions as used herein can be used interchangeably.

[0069] In one embodiment, the following terminology applies to the representation of geographic features in the geographic database 105.

[0070] "Node" - A point that terminates a link.

[0071] "Line segment" - A straight line connecting two points.

[0072] "Link" (or "edge") - A contiguous, non-branching string of one or more line segments terminating in a node at each end.

[0073] "Shape point" - A point along a link between two nodes (e.g., used to alter a shape of the link without defining new nodes).

[0074] "Oriented link" - A link that has a starting node (referred to as the "reference node") and an ending node (referred to as the "non reference node").

[0075] "Simple polygon" - An interior area of an outer boundary formed by a string of oriented links that begins and ends in one node. In one embodiment, a simple polygon does not cross itself.

[0076] "Polygon" - An area bounded by an outer boundary and none or at least one interior boundary (e.g., a hole or island). In one embodiment, a polygon is constructed from one outer simple polygon and none or at least one inner simple polygon. A polygon is simple if it just consists of one simple polygon, or complex if it has at least one inner simple polygon.

[0077] In one embodiment, the geographic database 105 follows certain conventions. For example, links do not cross themselves and do not cross each other except at a node. Also, there are no duplicated shape points, nodes, or links. Two links that connect each other have a common node. In the geographic database 105, overlapping geographic features are represented by overlapping polygons. When polygons overlap, the boundary of one polygon crosses the boundary of the other polygon. In the geographic database 105, the location at which the boundary of one polygon intersects they boundary of another polygon is represented by a node. In one embodiment, a node may be used to represent other locations along the boundary of a polygon than a location at which the boundary of the polygon intersects the boundary of another polygon. In one embodiment, a shape point is not used to represent a point at which the boundary of a polygon intersects the boundary of another polygon.

[0078] As shown, the geographic database 105 includes node data records 903, road segment or link data records 905, POI data records 907, ground control point data records 909, HD mapping data records 911, and indexes 913, for example. More, fewer or different data records can be provided. In one embodiment, additional data records (not shown) can include cartographic ("carto") data records, routing data, and maneuver data. In one embodiment, the indexes 913 may improve the speed of data retrieval operations in the geographic database 105. In one embodiment, the indexes 913 may be used to quickly locate data without having to search every row in the geographic database 105 every time it is accessed. For example, in one embodiment, the indexes 913 can be a spatial index of the polygon points associated with stored feature polygons.

[0079] In exemplary embodiments, the road segment data records 905 are links or segments representing roads, streets, or paths, as can be used in the calculated route or recorded route information for determination of one or more personalized routes. The node data records 903 are end points corresponding to the respective links or segments of the road segment data records 905. The road link data records 905 and the node data records 903 represent a road network, such as used by vehicles, cars, and/or other entities. Alternatively, the geographic database 105 can contain path segment and node data records or other data that represent pedestrian paths or areas in addition to or instead of the vehicle road record data, for example.

[0080] The road/link segments and nodes can be associated with attributes, such as functional class, a road elevation, a speed category, a presence or absence of road features, geographic coordinates, street names, address ranges, speed limits, turn restrictions at intersections, and other navigation related attributes, as well as POIs, such as gasoline stations, hotels, restaurants, museums, stadiums, offices, automobile dealerships, auto repair shops, buildings, stores, parks, etc. The geographic database 105 can include data about the POIs and their respective locations in the POI data records 907. The geographic database 105 can also include data about places, such as cities, towns, or other communities, and other geographic features, such as bodies of water, mountain ranges, etc. Such place or feature data can be part of the POI data records 907 or can be associated with POIs or POI data records 907 (such as a data point used for displaying or representing a position of a city).

[0081] In one embodiment, the geographic database 105 can also include ground control point data records 909 for storing the ground control point data, learnable map features, as well as other related data used according to the various embodiments described herein. In addition, the ground control point data records 909 can also store ground truth training and evaluation data, machine learning models, annotated observations, and/or any other data generated or used by the system 100 according to the various embodiments described herein. By way of example, the ground control point data records 909 can be associated with one or more of the node records 903, road segment records 905, and/or POI data records 907 to support localization or visual odometry based on the features stored therein and the corresponding estimated quality of the features. In this way, the records 909 can also be associated with or used to classify the characteristics or metadata of the corresponding records 903, 905, and/or 907.

[0082] In one embodiment, as discussed above, the HD mapping data records 911 model road surfaces and other map features to centimeter-level or better accuracy. The HD mapping data records 911 also include lane models that provide the precise lane geometry with lane boundaries, as well as rich attributes of the lane models. These rich attributes include, but are not limited to, lane traversal information, lane types, lane marking types, lane level speed limit information, and/or the like. In one embodiment, the HD mapping data records 911 are divided into spatial partitions of varying sizes to provide HD mapping data to vehicles 101 and other end user devices with near real-time speed without overloading the available resources of the vehicles 101 and/or devices (e.g., computational, memory, bandwidth, etc. resources).

[0083] In one embodiment, the HD mapping data records 911 are created from high-resolution 3D mesh or point-cloud data generated, for instance, from LiDAR-equipped vehicles. The 3D mesh or point-cloud data are processed to create 3D representations of a street or geographic environment at centimeter-level accuracy for storage in the HD mapping data records 911.

[0084] In one embodiment, the HD mapping data records 911 also include real-time sensor data collected from probe vehicles in the field. The real-time sensor data, for instance, integrates real-time traffic information, weather, and road conditions (e.g., potholes, road friction, road wear, etc.) with highly detailed 3D representations of street and geographic features to provide precise real-time also at centimeter-level accuracy. Other sensor data can include vehicle telemetry or operational data such as windshield wiper activation state, braking state, steering angle, accelerator position, and/or the like.

[0085] In one embodiment, the geographic database 105 can be maintained by the content provider 121 in association with the services platform 111 (e.g., a map developer). The map developer can collect geographic data to generate and enhance the geographic database 105. There can be different ways used by the map developer to collect data. These ways can include obtaining data from other sources, such as municipalities or respective geographic authorities. In addition, the map developer can employ field personnel to travel by vehicle (e.g., vehicle 101 and/or UE 115) along roads throughout the geographic region to observe features and/or record information about them, for example. Also, remote sensing, such as aerial or satellite photography, can be used.

[0086] The geographic database 105 can be a master geographic database stored in a format that facilitates updating, maintenance, and development. For example, the master geographic database or data in the master geographic database can be in an Oracle spatial format or other spatial format, such as for development or production purposes. The Oracle spatial format or development/production database can be compiled into a delivery format, such as a geographic data files (GDF) format. The data in the production and/or delivery formats can be compiled or further compiled to form geographic database products or databases, which can be used in end user navigation devices or systems.

[0087] For example, geographic data is compiled (such as into a platform specification format (PSF) format) to organize and/or configure the data for performing navigation-related functions and/or services, such as route calculation, route guidance, map display, speed calculation, distance and travel time functions, and other functions, by a navigation device, such as by a vehicle 101 or UE 115, for example. The navigation-related functions can correspond to vehicle navigation, pedestrian navigation, or other types of navigation. The compilation to produce the end user databases can be performed by a party or entity separate from the map developer. For example, a customer of the map developer, such as a navigation device developer or other end user device developer, can perform compilation on a received geographic database in a delivery format to produce one or more compiled navigation databases.

[0088] The processes described herein for determining ground control points from image data may be advantageously implemented via software, hardware (e.g., general processor, Digital Signal Processing (DSP) chip, an Application Specific Integrated Circuit (ASIC), Field Programmable Gate Arrays (FPGAs), etc.), firmware or a combination thereof. Such exemplary hardware for performing the described functions is detailed below.

[0089] FIG. 10 illustrates a computer system 1000 upon which an embodiment of the invention may be implemented. Computer system 1000 is programmed (e.g., via computer program code or instructions) to determine ground control points from image data as described herein and includes a communication mechanism such as a bus 1010 for passing information between other internal and external components of the computer system 1000. Information (also called data) is represented as a physical expression of a measurable phenomenon, typically electric voltages, but including, in other embodiments, such phenomena as magnetic, electromagnetic, pressure, chemical, biological, molecular, atomic, sub-atomic and quantum interactions. For example, north and south magnetic fields, or a zero and non-zero electric voltage, represent two states (0, 1) of a binary digit (bit). Other phenomena can represent digits of a higher base. A superposition of multiple simultaneous quantum states before measurement represents a quantum bit (qubit). A sequence of one or more digits constitutes digital data that is used to represent a number or code for a character. In some embodiments, information called analog data is represented by a near continuum of measurable values within a particular range.

[0090] A bus 1010 includes one or more parallel conductors of information so that information is transferred quickly among devices coupled to the bus 1010. One or more processors 1002 for processing information are coupled with the bus 1010.

[0091] A processor 1002 performs a set of operations on information as specified by computer program code related to determining ground control points from image data. The computer program code is a set of instructions or statements providing instructions for the operation of the processor and/or the computer system to perform specified functions. The code, for example, may be written in a computer programming language that is compiled into a native instruction set of the processor. The code may also be written directly using the native instruction set (e.g., machine language). The set of operations include bringing information in from the bus 1010 and placing information on the bus 1010. The set of operations also typically include comparing two or more units of information, shifting positions of units of information, and combining two or more units of information, such as by addition or multiplication or logical operations like OR, exclusive OR (XOR), and AND. Each operation of the set of operations that can be performed by the processor is represented to the processor by information called instructions, such as an operation code of one or more digits. A sequence of operations to be executed by the processor 1002, such as a sequence of operation codes, constitute processor instructions, also called computer system instructions or, simply, computer instructions. Processors may be implemented as mechanical, electrical, magnetic, optical, chemical or quantum components, among others, alone or in combination.

[0092] Computer system 1000 also includes a memory 1004 coupled to bus 1010. The memory 1004, such as a random access memory (RAM) or other dynamic storage device, stores information including processor instructions for determining ground control points from image data. Dynamic memory allows information stored therein to be changed by the computer system 1000. RAM allows a unit of information stored at a location called a memory address to be stored and retrieved independently of information at neighboring addresses. The memory 1004 is also used by the processor 1002 to store temporary values during execution of processor instructions. The computer system 1000 also includes a read only memory (ROM) 1006 or other static storage device coupled to the bus 1010 for storing static information, including instructions, that is not changed by the computer system 1000. Some memory is composed of volatile storage that loses the information stored thereon when power is lost. Also coupled to bus 1010 is a non-volatile (persistent) storage device 1008, such as a magnetic disk, optical disk or flash card, for storing information, including instructions, that persists even when the computer system 1000 is turned off or otherwise loses power.

[0093] Information, including instructions for determining ground control points from image data, is provided to the bus 1010 for use by the processor from an external input device 1012, such as a keyboard containing alphanumeric keys operated by a human user, or a sensor. A sensor detects conditions in its vicinity and transforms those detections into physical expression compatible with the measurable phenomenon used to represent information in computer system 1000. Other external devices coupled to bus 1010, used primarily for interacting with humans, include a display device 1014, such as a cathode ray tube (CRT) or a liquid crystal display (LCD), or plasma screen or printer for presenting text or images, and a pointing device 1016, such as a mouse or a trackball or cursor direction keys, or motion sensor, for controlling a position of a small cursor image presented on the display 1014 and issuing commands associated with graphical elements presented on the display 1014. In some embodiments, for example, in embodiments in which the computer system 1000 performs all functions automatically without human input, one or more of external input device 1012, display device 1014 and pointing device 1016 is omitted.

[0094] In the illustrated embodiment, special purpose hardware, such as an application specific integrated circuit (ASIC) 1020, is coupled to bus 1010. The special purpose hardware is configured to perform operations not performed by processor 1002 quickly enough for special purposes. Examples of application specific ICs include graphics accelerator cards for generating images for display 1014, cryptographic boards for encrypting and decrypting messages sent over a network, speech recognition, and interfaces to special external devices, such as robotic arms and medical scanning equipment that repeatedly perform some complex sequence of operations that are more efficiently implemented in hardware.

[0095] Computer system 1000 also includes one or more instances of a communications interface 1070 coupled to bus 1010. Communication interface 1070 provides a one-way or two-way communication coupling to a variety of external devices that operate with their own processors, such as printers, scanners and external disks. In general the coupling is with a network link 1078 that is connected to a local network 1080 to which a variety of external devices with their own processors are connected. For example, communication interface 1070 may be a parallel port or a serial port or a universal serial bus (USB) port on a personal computer. In some embodiments, communications interface 1070 is an integrated services digital network (ISDN) card or a digital subscriber line (DSL) card or a telephone modem that provides an information communication connection to a corresponding type of telephone line. In some embodiments, a communication interface 1070 is a cable modem that converts signals on bus 1010 into signals for a communication connection over a coaxial cable or into optical signals for a communication connection over a fiber optic cable. As another example, communications interface 1070 may be a local area network (LAN) card to provide a data communication connection to a compatible LAN, such as Ethernet. Wireless links may also be implemented. For wireless links, the communications interface 1070 sends or receives or both sends and receives electrical, acoustic or electromagnetic signals, including infrared and optical signals, that carry information streams, such as digital data. For example, in wireless handheld devices, such as mobile telephones like cell phones, the communications interface 1070 includes a radio band electromagnetic transmitter and receiver called a radio transceiver. In certain embodiments, the communications interface 1070 enables connection to the communication network 119 for determining ground control points from image data.

[0096] The term computer-readable medium is used herein to refer to any medium that participates in providing information to processor 1002, including instructions for execution. Such a medium may take many forms, including, but not limited to, non-volatile media, volatile media and transmission media. Non-volatile media include, for example, optical or magnetic disks, such as storage device 1008. Volatile media include, for example, dynamic memory 1004. Transmission media include, for example, coaxial cables, copper wire, fiber optic cables, and carrier waves that travel through space without wires or cables, such as acoustic waves and electromagnetic waves, including radio, optical and infrared waves. Signals include man-made transient variations in amplitude, frequency, phase, polarization or other physical properties transmitted through the transmission media. Common forms of computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, CDRW, DVD, any other optical medium, punch cards, paper tape, optical mark sheets, any other physical medium with patterns of holes or other optically recognizable indicia, a RAM, a PROM, an EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave, or any other medium from which a computer can read.

[0097] FIG. 11 illustrates a chip set 1100 upon which an embodiment of the invention may be implemented. Chip set 1100 is programmed to determine ground control points from image data as described herein and includes, for instance, the processor and memory components described with respect to FIG. 10 incorporated in one or more physical packages (e.g., chips). By way of example, a physical package includes an arrangement of one or more materials, components, and/or wires on a structural assembly (e.g., a baseboard) to provide one or more characteristics such as physical strength, conservation of size, and/or limitation of electrical interaction. It is contemplated that in certain embodiments the chip set can be implemented in a single chip.

[0098] In one embodiment, the chip set 1100 includes a communication mechanism such as a bus 1101 for passing information among the components of the chip set 1100. A processor 1103 has connectivity to the bus 1101 to execute instructions and process information stored in, for example, a memory 1105. The processor 1103 may include one or more processing cores with each core configured to perform independently. A multi-core processor enables multiprocessing within a single physical package. Examples of a multi-core processor include two, four, eight, or greater numbers of processing cores. Alternatively or in addition, the processor 1103 may include one or more microprocessors configured in tandem via the bus 1101 to enable independent execution of instructions, pipelining, and multithreading. The processor 1103 may also be accompanied with one or more specialized components to perform certain processing functions and tasks such as one or more digital signal processors (DSP) 1107, or one or more application-specific integrated circuits (ASIC) 1109. A DSP 1107 typically is configured to process real-world signals (e.g., sound) in real time independently of the processor 1103. Similarly, an ASIC 1109 can be configured to performed specialized functions not easily performed by a general purposed processor. Other specialized components to aid in performing the inventive functions described herein include one or more field programmable gate arrays (FPGA) (not shown), one or more controllers (not shown), or one or more other special-purpose computer chips.

[0099] The processor 1103 and accompanying components have connectivity to the memory 1105 via the bus 1101. The memory 1105 includes both dynamic memory (e.g., RAM, magnetic disk, writable optical disk, etc.) and static memory (e.g., ROM, CD-ROM, etc.) for storing executable instructions that when executed perform the inventive steps described herein to determine ground control points from image data. The memory 1105 also stores the data associated with or generated by the execution of the inventive steps.

[0100] FIG. 12 is a diagram of exemplary components of a mobile terminal (e.g., UE 115 or embedded component of the vehicle 101) capable of operating in the system of FIG. 1, according to one embodiment. Generally, a radio receiver is often defined in terms of front-end and back-end characteristics. The front-end of the receiver encompasses all of the Radio Frequency (RF) circuitry whereas the back-end encompasses all of the base-band processing circuitry. Pertinent internal components of the telephone include a Main Control Unit (MCU) 1203, a Digital Signal Processor (DSP) 1205, and a receiver/transmitter unit including a microphone gain control unit and a speaker gain control unit. A main display unit 1207 provides a display to the user in support of various applications and mobile station functions that offer automatic contact matching. An audio function circuitry 1209 includes a microphone 1211 and microphone amplifier that amplifies the speech signal output from the microphone 1211. The amplified speech signal output from the microphone 1211 is fed to a coder/decoder (CODEC) 1213.

[0101] A radio section 1215 amplifies power and converts frequency in order to communicate with a base station, which is included in a mobile communication system, via antenna 1217. The power amplifier (PA) 1219 and the transmitter/modulation circuitry are operationally responsive to the MCU 1203, with an output from the PA 1219 coupled to the duplexer 1221 or circulator or antenna switch, as known in the art. The PA 1219 also couples to a battery interface and power control unit 1220.

[0102] In use, a user of mobile station 1201 speaks into the microphone 1211 and his or her voice along with any detected background noise is converted into an analog voltage. The analog voltage is then converted into a digital signal through the Analog to Digital Converter (ADC) 1223. The control unit 1203 routes the digital signal into the DSP 1205 for processing therein, such as speech encoding, channel encoding, encrypting, and interleaving. In one embodiment, the processed voice signals are encoded, by units not separately shown, using a cellular transmission protocol such as global evolution (EDGE), general packet radio service (GPRS), global system for mobile communications (GSM), Internet protocol multimedia subsystem (IMS), universal mobile telecommunications system (UMTS), etc., as well as any other suitable wireless medium, e.g., microwave access (WiMAX), Long Term Evolution (LTE) networks, code division multiple access (CDMA), wireless fidelity (WiFi), satellite, and the like.

[0103] The encoded signals are then routed to an equalizer 1225 for compensation of any frequency-dependent impairments that occur during transmission though the air such as phase and amplitude distortion. After equalizing the bit stream, the modulator 1227 combines the signal with a RF signal generated in the RF interface 1229. The modulator 1227 generates a sine wave by way of frequency or phase modulation. In order to prepare the signal for transmission, an up-converter 1231 combines the sine wave output from the modulator 1227 with another sine wave generated by a synthesizer 1233 to achieve the desired frequency of transmission. The signal is then sent through a PA 1219 to increase the signal to an appropriate power level. In practical systems, the PA 1219 acts as a variable gain amplifier whose gain is controlled by the DSP 1205 from information received from a network base station. The signal is then filtered within the duplexer 1221 and optionally sent to an antenna coupler 1235 to match impedances to provide maximum power transfer. Finally, the signal is transmitted via antenna 1217 to a local base station. An automatic gain control (AGC) can be supplied to control the gain of the final stages of the receiver. The signals may be forwarded from there to a remote telephone which may be another cellular telephone, other mobile phone or a land-line connected to a Public Switched Telephone Network (PSTN), or other telephony networks.

[0104] Voice signals transmitted to the mobile station 1201 are received via antenna 1217 and immediately amplified by a low noise amplifier (LNA) 1237. A down-converter 1239 lowers the carrier frequency while the demodulator 1241 strips away the RF leaving only a digital bit stream. The signal then goes through the equalizer 1225 and is processed by the DSP 1205. A Digital to Analog Converter (DAC) 1243 converts the signal and the resulting output is transmitted to the user through the speaker 1245, all under control of a Main Control Unit (MCU) 1203-which can be implemented as a Central Processing Unit (CPU) (not shown).

[0105] The MCU 1203 receives various signals including input signals from the keyboard 1247. The keyboard 1247 and/or the MCU 1203 in combination with other user input components (e.g., the microphone 1211) comprise a user interface circuitry for managing user input. The MCU 1203 runs a user interface software to facilitate user control of at least some functions of the mobile station 1201 to determine ground control points from image data. The MCU 1203 also delivers a display command and a switch command to the display 1207 and to the speech output switching controller, respectively. Further, the MCU 1203 exchanges information with the DSP 1205 and can access an optionally incorporated SIM card 1249 and a memory 1251. In addition, the MCU 1203 executes various control functions required of the station. The DSP 1205 may, depending upon the implementation, perform any of a variety of conventional digital processing functions on the voice signals. Additionally, DSP 1205 determines the background noise level of the local environment from the signals detected by microphone 1211 and sets the gain of microphone 1211 to a level selected to compensate for the natural tendency of the user of the mobile station 1201.

[0106] The CODEC 1213 includes the ADC 1223 and DAC 1243. The memory 1251 stores various data including call incoming tone data and is capable of storing other data including music data received via, e.g., the global Internet. The software module could reside in RAM memory, flash memory, registers, or any other form of writable computer-readable storage medium known in the art including non-transitory computer-readable storage medium. For example, the memory device 1251 may be, but not limited to, a single memory, CD, DVD, ROM, RAM, EEPROM, optical storage, or any other non-volatile or non-transitory storage medium capable of storing digital data.

[0107] An optionally incorporated SIM card 1249 carries, for instance, important information, such as the cellular phone number, the carrier supplying service, subscription details, and security information. The SIM card 1249 serves primarily to identify the mobile station 1201 on a radio network. The card 1249 also contains a memory for storing a personal telephone number registry, text messages, and user specific mobile station settings.

[0108] While the invention has been described in connection with a number of embodiments and implementations, the invention is not so limited but covers various obvious modifications and equivalent arrangements, which fall within the purview of the appended claims. Although features of the invention are expressed in certain combinations among the claims, it is contemplated that these features can be arranged in any combination and order.


Claims

1. A computer-implemented method for determining a ground control point from image data comprising:

selecting a feature as the ground control point based determining that the feature meets one or more properties for classification as a machine learnable feature;

retrieving a plurality of ground truth images depicting the feature, wherein the plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images; and

training a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.


 
2. The method of claim 1, wherein the feature is selected based on determining that the one or more properties indicate that the feature is uniquely identifiable from among other features.
 
3. The method of claim 1, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature has a spatial sparsity that meets a sparsity criterion.
 
4. The method of claim 1, wherein the feature is selected based on determining that the one or more properties indicate that the feature is applicable to a plurality of different geographic regions.
 
5. The method of claim 1, wherein the feature is selected from a category of curvilinear geometry features intersection features including a gore point, a cross-walk corner, a boundary of a land boundary and a limit line, or a combination thereof.
 
6. The method of claim 1, wherein feature is a visible feature of a roadway intersection that is visible from a top-down imagery perspective.
 
7. The method of claim 1, wherein the training features for the machine learning model include a pixel location of the intersection feature in said each of the plurality of ground truth images, a camera position of said each of the plurality of ground truth images, an orientation of said each of the plurality of ground truth images, a focal length of said each of the plurality of ground truth images, or a combination thereof.
 
8. The method of claim 1, further comprising:

processing a plurality of input images using the trained machine learning model to identify that the ground control point is present in the plurality of input images; and

triangulating a real-world location of the ground control point based on a pixel correspondence of the ground control point among the plurality of input images.


 
9. The method of claim 8, wherein the real-world location is a three-dimensional location of the ground control point.
 
10. The method of claim 1, wherein the machine learning model is further trained to calculate an uncertainty associated with the predicted location based on a characteristic of said each of the plurality of images, a respective source of said each of the plurality of images, or a combination.
 
11. An apparatus for determining a ground control point from image data comprising:

at least one processor; and

at least one memory including computer program code for one or more programs,

the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following,

select a feature as the ground control point based determining that the feature meets one or more properties for classification as a machine learnable feature;

retrieve a plurality of ground truth images depicting the feature, wherein the plurality of ground truth images is labeled with known pixel location data of the feature as respectively depicted in each of the plurality of ground truth images; and

train a machine learning model using the plurality of ground truth images to identify predicted pixel location data of the ground control point as depicted in an input image.


 
12. The apparatus of claim 11, wherein the feature is selected based on determining that the one or more properties indicate that the feature is uniquely identifiable from among other features.
 
13. The apparatus of claim 11, wherein the feature is selected based on determining that the one or more properties indicate that the intersection feature has a spatial sparsity that meets a sparsity criterion.
 
14. The apparatus of claim 11, wherein the feature is selected based on determining that the one or more properties indicate that the feature is applicable to a plurality of different geographic regions.
 
15. The apparatus of claim 11, wherein the feature is selected from a category of curvilinear geometry features intersection features including a gore point, a cross-walk corner, a boundary of a land boundary and a limit line, or a combination thereof.
 




Drawing














































Search report









Search report