A method, apparatus, medium, and equipment for attitude recognition of large obstacles.

By acquiring obstacle image data, performing region segmentation and feature point matching, and calculating the obstacle's orientation angle, the shortcomings of automatic parking systems in recognizing the posture of large obstacles are addressed, thus improving parking recognition capabilities.

CN118506325BActive Publication Date: 2024-11-01CHINA AUTOMOBILE ZHILIAN TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410957951.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-01
Estimated Expiration
2044-07-17

AI Technical Summary

Technical Problem

Existing automatic parking systems cannot effectively recognize the posture of large obstacles at a distance, which may lead to the risk of parking trajectories being occupied or non-parking spaces being abnormally released.

Method used

By acquiring image data of the target obstacle, region segmentation is performed to identify bounding boxes, feature points are matched, the orientation angle of the target obstacle is calculated, and the relative orientation between the obstacle and the vehicle is determined by combining the orientation of the target vehicle.

Benefits of technology

It improves the ability to recognize large obstacles, ensuring the effective operation of the automatic parking system in different scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118506325B_ABST
    Figure CN118506325B_ABST
Patent Text Reader

Abstract

The present application discloses a posture recognition method, device, medium and equipment for large-scale obstacle, which obtains a bounding box of the target obstacle by acquiring image data containing the target obstacle and performing region segmentation; matches the feature points of the target obstacle in the bounding box; selects multiple target feature points at the same height and calculates the orientation angle of the target obstacle; that is, obtains a bounding box of the target obstacle by acquiring an image of the target obstacle and identifying it, matches the feature points of the target obstacle in the bounding box, determines the orientation of the target obstacle based on multiple feature points at the same height of the target obstacle, and calculates the orientation angle between the target obstacle and the target vehicle in combination with the orientation of the target vehicle, so that the relative orientation between the target obstacle and the target vehicle at different distances can be calculated according to the feature points at different heights in the image of the target obstacle, so as to be applicable to the recognition of large obstacles in various scenarios, thereby improving the recognition capability of automatic parking.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of automatic obstacle recognition, and in particular to a method, device, medium and equipment for posture recognition of large obstacles. Background Art

[0002] Automatic parking refers to the automatic parking of a car, which is an important task in intelligent driving. Some large obstacles (such as trucks) are huge and have different shapes. If large obstacles cannot be effectively identified, especially when searching for parking spaces, the trajectory of available parking spaces will be occupied, or the unavailable parking spaces (trajectory is occupied) will be abnormally released, resulting in the risk of parking interruption or failure to identify necessary blind spots.

[0003] The current automatic parking system mainly relies on the vehicle passing and scanning the corresponding parking space, and uses ultrasonic radar for auxiliary judgment. However, the detection distance of ultrasonic radar is limited, and it cannot effectively detect obstacles at a long distance, and its ability to recognize the posture of spatial obstacles is limited. Therefore, a method that can be applied to long-distance and short-distance recognition of large obstacles is needed. Summary of the invention

[0004] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a posture recognition method, device, medium and equipment for large obstacles.

[0005] According to one aspect of the present application, a large obstacle posture recognition method is provided, which is applied to a target vehicle, and the large obstacle posture recognition method includes: acquiring image data containing the target obstacle; performing region segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents a minimum geometric frame containing the target obstacle; according to feature information of the target obstacle, matching feature points of the target obstacle within the bounding box; wherein the feature points represent landmark points at specific locations on the target obstacle; selecting multiple target feature points at the same height among the feature points, and calculating an orientation angle of the target obstacle; wherein the orientation angle represents an angle between the orientations of the target obstacle and the target vehicle.

[0006] In one embodiment, selecting multiple target feature points at the same height among the feature points and calculating the orientation angle of the target obstacle includes: if there are multiple ground feature points among the feature points, calculating the orientation angle of the target obstacle based on the multiple ground feature points; wherein the ground feature point is a feature point where the target obstacle contacts the ground.

[0007] In one embodiment, if there are multiple ground feature points among the feature points, calculating the orientation angle of the target obstacle based on the multiple ground feature points includes: if the number of the ground feature points is greater than 2, selecting the two ground feature points with the farthest relative distance to calculate the orientation angle of the target obstacle.

[0008] In one embodiment, selecting multiple target feature points at the same height among the feature points and calculating the orientation angle of the target obstacle includes: if there are multiple central feature points at the same height among the feature points, calculating the orientation angle of the target obstacle based on the multiple central feature points; wherein the central feature point is the geometric center point of the target obstacle.

[0009] In one embodiment, selecting multiple target feature points at the same height among the feature points and calculating the orientation angle of the target obstacle includes: if there are multiple top feature points among the feature points, calculating the orientation angle of the target obstacle based on the multiple top feature points; wherein the top feature point is the top feature point of the target obstacle.

[0010] In one embodiment, if there are multiple top feature points among the feature points, calculating the orientation angle of the target obstacle based on the multiple top feature points includes: if the number of the top feature points is greater than 2, selecting the two top feature points with the farthest relative distance to calculate the orientation angle of the target obstacle.

[0011] In one embodiment, selecting multiple target feature points at the same height among the feature points and calculating the orientation angle of the target obstacle includes: if multiple orientation angles are calculated, eliminating abnormal values ​​among the multiple orientation angles to obtain the final orientation angle of the target obstacle.

[0012] According to another aspect of the present application, a posture recognition device for a large obstacle is provided, which is arranged on a target vehicle, and the posture recognition device for a large obstacle includes: an image acquisition module, which is used to acquire image data containing a target obstacle; a boundary division module, which is used to perform region segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents a minimum geometric frame containing the target obstacle; a feature point matching module, which is used to match feature points of the target obstacle within the bounding box according to feature information of the target obstacle; wherein the feature points represent landmark points at specific positions on the target obstacle; and a heading angle calculation module, which is used to select multiple target feature points at the same height among the feature points, and calculate the heading angle of the target obstacle; wherein the heading angle represents the angle between the headings of the target obstacle and the target vehicle.

[0013] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute any of the above-mentioned large obstacle posture recognition methods.

[0014] According to another aspect of the present application, an electronic device is provided, comprising: a processor; a memory for storing instructions executable by the processor; the processor is used to execute any of the above-mentioned large obstacle posture recognition methods.

[0015] The present application provides a large obstacle posture recognition method, device, medium and equipment, which obtain image data containing a target obstacle; perform regional segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents the minimum geometric frame containing the target obstacle; according to the feature information of the target obstacle, match the feature points of the target obstacle in the bounding box; wherein the feature points represent the landmark points at a specific position on the target obstacle; select multiple target feature points at the same height from the feature points to calculate the orientation angle of the target obstacle; wherein the orientation angle represents the angle between the orientations of the target obstacle and the target vehicle; that is, by obtaining an image of the target obstacle and identifying the bounding box of the target obstacle, matching the feature points of the target obstacle in the bounding box, determining the orientation of the target obstacle based on multiple feature points at the same height of the target obstacle, and calculating the orientation angle between the target obstacle and the target vehicle in combination with the orientation of the target vehicle, so that the relative orientations between the target obstacle and the target vehicle at different distances can be calculated according to the feature points at different heights in the image of the target obstacle, so as to be applicable to the recognition of large obstacles in various scenarios, thereby improving the recognition capability of automatic parking. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 It is a flowchart of a method for posture recognition of a large obstacle provided by an exemplary embodiment of the present application.

[0018] Figure 2 It is a flowchart of a method for calculating the orientation angle of a large obstacle provided by an exemplary embodiment of the present application.

[0019] Figure 3 It is a structural schematic diagram of a posture recognition device for a large obstacle provided by an exemplary embodiment of the present application.

[0020] Figure 4 It is a structural diagram of an electronic device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] Figure 1 FIG. 1 is a flow chart of a method for recognizing a large obstacle according to an exemplary embodiment of the present invention. The method for recognizing a large obstacle is applied to a target vehicle, such as Figure 1 As shown, the large obstacle posture recognition method includes the following steps:

[0023] Step 110: Acquire image data containing the target obstacle.

[0024] During the parking process of the target vehicle, the on-board camera installed on the target vehicle is used to obtain image data around the target vehicle in real time. The on-board camera has a large field of view. When there are large obstacles around the target vehicle, the image data will contain the target obstacle (i.e., the large obstacle).

[0025] Step 120: Perform region segmentation on the image data to obtain a bounding box of the target obstacle.

[0026] Among them, the bounding box of the target obstacle represents the minimum geometric bounding box that contains the target obstacle. After acquiring the image data containing the target obstacle, the image data is segmented to segment out the minimum bounding box containing the target obstacle, thereby reducing the interference of the background image, while also reducing the amount of calculated data, and improving calculation efficiency and response efficiency. Specifically, the present application can identify the boundary line between the target obstacle image and the background image in the image data through a boundary recognition method, thereby segmenting the target obstacle, and generating a bounding box (such as a rectangle, etc.) of the target obstacle based on the boundary line of the target obstacle. The present application abstracts the target obstacle as a geometric bounding box, which can simplify the subsequent calculation method of the orientation angle of the target obstacle.

[0027] Step 130: Matching feature points of the target obstacle within the bounding box according to the feature information of the target obstacle.

[0028] Among them, the feature point represents a landmark point at a specific position on the target obstacle. After obtaining the bounding box of the target obstacle, the present application matches the feature points of the target obstacle within the bounding box based on the feature information and bounding box of the target obstacle. Since the distance between the target obstacle and the target vehicle may be far or close, when the distance between the target obstacle and the target vehicle is far, the on-board camera can obtain image data of the lower position of the target obstacle (for example, the contact point with the ground), and when the distance between the target obstacle and the target vehicle is close, the on-board camera may only be able to obtain image data of the higher position of the target obstacle. Therefore, the present application divides the feature points of the target obstacle into the ground point (the contact point between the target obstacle and the ground), the middle point (the point located in the middle of the target obstacle) and the upper vertex (the point located at the top of the target obstacle).

[0029] Step 140: Select multiple target feature points at the same height among the feature points, and calculate the orientation angle of the target obstacle.

[0030] The orientation angle represents the angle between the orientation of the target obstacle and the orientation of the target vehicle. After matching multiple feature points of the target obstacle, multiple feature points at the same height are selected to calculate the orientation angle of the target obstacle. Specifically, the connecting line of multiple feature points at the same height is used as the orientation of the target obstacle. Combined with the orientation of the target vehicle, the angle between the orientation of the target obstacle and the orientation of the target vehicle is calculated, that is, the orientation angle of the target obstacle (the deviation angle relative to the orientation of the target vehicle) is obtained.

[0031] The present application provides a large obstacle posture recognition method, which includes acquiring image data containing a target obstacle; performing regional segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents a minimum geometric frame containing the target obstacle; matching feature points of the target obstacle within the bounding box according to feature information of the target obstacle; wherein the feature point represents a landmark point at a specific position on the target obstacle; selecting multiple target feature points at the same height among the feature points to calculate the orientation angle of the target obstacle; wherein the orientation angle represents the angle between the orientations of the target obstacle and a target vehicle; that is, obtaining a bounding box of the target obstacle by acquiring an image of the target obstacle and identifying it, matching feature points of the target obstacle within the bounding box, determining the orientation of the target obstacle based on multiple feature points at the same height of the target obstacle, and calculating the orientation angle between the target obstacle and the target vehicle in combination with the orientation of the target vehicle, so that the relative orientation between the target obstacle and the target vehicle at different distances can be calculated according to feature points at different heights in the image of the target obstacle, so as to be applicable to the recognition of large obstacles in various scenarios, thereby improving the recognition capability of automatic parking.

[0032] In one embodiment, the specific implementation method of the above step 140 can be: if there are multiple ground feature points among the feature points, the orientation angle of the target obstacle is calculated based on the multiple ground feature points; wherein the ground feature point is the feature point where the target obstacle contacts the ground.

[0033] Since the ground is a fixed reference object and the contact point between the ground and the target obstacle is relatively easy to determine, if the number of ground feature points of the target obstacle in the image data is greater than or equal to 2, the line connecting the ground feature points can be given priority as the orientation of the target obstacle, and the orientation angle of the target obstacle can be calculated based on this. Specifically, the aperture point of the vehicle-mounted camera is used as the origin to establish a spatial physical coordinate system, and the pixel coordinate system is established with the pixel where the aperture point of the vehicle-mounted camera falls on the generated image as the origin, where the physical coordinates of the feature points are and pixel coordinates The relationship between them is:

[0034] ,

[0035] ;

[0036] in, and The vehicle-mounted camera is X and Y The focal length in the direction, and are the pixel coordinates of the point where the optical axis falls in the plane image, K is the matrix of the camera's intrinsic parameters.

[0037] Note two grounding feature points The coordinates of , , then the direction angle of the target obstacle .

[0038] In one embodiment, the specific implementation of step 140 may be: if the number of ground feature points is greater than 2, two ground feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

[0039] Since there is more than one contact point between a wheel and the ground, and a large vehicle may have more than two wheels on one side, the target obstacle in the acquired image data has more than two grounding feature points, and a single wheel also has multiple contact points with the ground. In order to maximize the calculation accuracy of the orientation angle of the target obstacle, the present application selects the line connecting the two farthest grounding feature points (usually grounding feature points on different wheels) as the orientation of the target obstacle. Preferably, after obtaining multiple grounding feature points, the present application can obtain the orientation of the target obstacle based on the fitting of the multiple grounding feature points.

[0040] Since the wheels of the target obstacle may not be aligned when parked, there is a deviation between the line connecting the ground feature points and the actual orientation of the target obstacle. The present application can determine whether the wheels of the target obstacle are aligned by comparing the degree of deviation between the line connecting the ground feature points and the edge line of the bounding box of the target obstacle. Specifically, if the angle between the line connecting the ground feature points and the edge line of the bounding box of the target obstacle is large, it means that the degree of deviation between the line connecting the ground feature points and the edge line of the bounding box of the target obstacle is large. At this time, it can be determined that the wheels of the target obstacle are not aligned. Otherwise, it is determined that the wheels of the target obstacle are aligned.

[0041] In one embodiment, the specific implementation method of the above step 140 can be: if there are multiple central feature points at the same height among the feature points, the orientation angle of the target obstacle is calculated based on the multiple central feature points; wherein the central feature point is the geometric center point of the target obstacle.

[0042] If the number of central feature points of the target obstacle in the image data is greater than or equal to 2, the line connecting the central feature points (such as the center points of the front and rear axles of the vehicle) can be used as the orientation of the target obstacle, and the orientation angle of the target obstacle can be calculated based on this. Specifically, the pixel coordinates of the central feature points are converted into 3D camera coordinates:

[0043] ;

[0044] in, is the pixel coordinate of the central feature point, is the rotation matrix, is the translation vector. The rotation matrix and the translation vector represent the transformation relationship between the vehicle coordinate system and the image coordinate system. K is the internal parameter matrix of the vehicle-mounted camera, is the 3D coordinate of the center feature point in the vehicle coordinate system, is the spatial height coordinate of the central feature point.

[0045] Since the image data does not contain height information, for height information, if the field of view of the vehicle camera contains the ground feature point of the target obstacle, this application will complete the information based on the ground plane to restore the height coordinates of the center feature point in the vehicle coordinate system. Specifically, the height coordinates of the center feature point The calculation formula is:

[0046] ;

[0047] in, It is the opposite of the height from the center point of the vehicle's rear axle to the ground.

[0048] If the distance between the target obstacle and the target vehicle is relatively narrow, there will be no ground feature points of the target obstacle in the field of view of the vehicle camera, and the height coordinates of the center feature point cannot be calculated based on the ground. In this case, the height coordinates of the center feature point in the vehicle coordinate system are calculated based on the front and rear height feature points of the front and rear wheels. ,in, is the corresponding vehicle in the camera coordinate system X, Y, Z The translation amount of the central feature point in the vehicle coordinate system is ( , , ), its coordinates in the intermediate coordinate system ( , , )for:

[0049] ;

[0050] The actual height of the central feature point in the camera coordinate system , the coordinate calculation formula of the central feature point in the intermediate coordinate system is:

[0051] ,

[0052] ,

[0053] ;

[0054] in,( , , 1) is the assumption When it is 1, the coordinates of the central feature point in the camera coordinate system are calculated. ( , , ) is the coordinate of the central feature point in the intermediate coordinate system,

[0055] Correspondingly, the calculation formula for the heading angle is:

[0056] ;

[0057] in, , are the heights of the two selected center feature points in the camera coordinate system; ( , ,1)、( , , 1) are the two selected central feature points in the assumption The coordinates in the intermediate coordinate system when is 1.

[0058] Since the front and rear wheel feature points have the same height in the vehicle coordinate system of the target vehicle, that is, , then the above formula can be further simplified to:

[0059] ,

[0060] ,

[0061] ;

[0062] in, , They are the coordinates of the selected central feature point in the pixel coordinate system.

[0063] If the number of central feature points is greater than 2, the two central feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle. Since the middle of the target obstacle includes multiple feature points, there are more than two central feature points of the target obstacle in the acquired image data. In order to maximize the calculation accuracy of the orientation angle of the target obstacle, the present application selects the line connecting the two central feature points with the farthest distance as the orientation of the target obstacle. Preferably, after obtaining multiple central feature points, the present application can obtain the orientation of the target obstacle based on the fitting of the multiple central feature points.

[0064] In one embodiment, the specific implementation method of the above step 140 can be: if there are multiple top feature points in the feature points, the orientation angle of the target obstacle is calculated according to the multiple top feature points; wherein the top feature point is the top feature point of the target obstacle.

[0065] If the number of the top feature points of the target obstacle in the image data is greater than or equal to 2, the line connecting the top feature points can be used as the orientation of the target obstacle, and the orientation angle of the target obstacle can be calculated based on this.

[0066] In one embodiment, the specific implementation of step 140 may be: if the number of upper feature points is greater than 2, two upper feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

[0067] Since the top of the target obstacle includes multiple feature points, there are more than two top feature points of the target obstacle in the acquired image data. In order to maximize the calculation accuracy of the orientation angle of the target obstacle, the present application selects the line connecting the two top feature points with the longest distance as the orientation of the target obstacle. Preferably, after obtaining multiple top feature points, the present application can obtain the orientation of the target obstacle based on the fitting of the multiple top feature points.

[0068] In one embodiment, the specific implementation of step 140 may be: if multiple orientation angles are calculated, then outliers among the multiple orientation angles are eliminated to obtain a final orientation angle of the target obstacle.

[0069] After calculating multiple orientation angles based on the ground feature point, the center feature point, and the top feature point, the present application uses the multiple orientation angles to verify each other to eliminate abnormal values, such as orientation angles that are significantly different from other orientation angles, thereby improving the calculation accuracy of the orientation angle of the target obstacle. Preferably, if Figure 2 As shown, the present application determines the final orientation angle of the target obstacle according to the priority of the ground feature point, the top feature point and the center feature point. Since the ground feature point is referenced to the ground, it is the easiest to obtain and has the highest accuracy, and the top feature point is referenced to the boundary of the target obstacle, it is easier to obtain and has a higher accuracy. Therefore, the present application sets the priority to select the feature points with higher accuracy as much as possible to calculate the orientation angle of the target obstacle.

[0070] Figure 3 1 is a schematic diagram of a large obstacle posture recognition device provided by an exemplary embodiment of the present application. The large obstacle posture recognition device is arranged on a target vehicle, such as Figure 3 As shown, the posture recognition device 30 for large obstacles includes: an image acquisition module 31, which is used to acquire image data containing a target obstacle; a boundary division module 32, which is used to perform region segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents the minimum geometric frame containing the target obstacle; a feature point matching module 33, which is used to match the feature points of the target obstacle within the bounding box according to the feature information of the target obstacle; wherein the feature points represent landmark points at specific positions on the target obstacle; and a heading angle calculation module 34, which is used to select multiple target feature points at the same height among the feature points and calculate the heading angle of the target obstacle; wherein the heading angle represents the angle between the headings of the target obstacle and the target vehicle.

[0071] The present application provides a posture recognition device for a large obstacle, which obtains image data containing a target obstacle through an image acquisition module 31; a boundary division module 32 performs regional segmentation on the image data to obtain a boundary box of the target obstacle; wherein the boundary box of the target obstacle represents the minimum geometric frame containing the target obstacle; a feature point matching module 33 matches the feature points of the target obstacle within the boundary box according to the feature information of the target obstacle; wherein the feature points represent landmark points at specific locations on the target obstacle; a heading angle calculation module 34 selects multiple target feature points at the same height among the feature points to calculate the heading angle of the target obstacle; wherein In the figure, the orientation angle represents the angle between the orientations of the target obstacle and the target vehicle; that is, the bounding box of the target obstacle is obtained by acquiring the image of the target obstacle and identifying it, the feature points of the target obstacle are matched within the bounding box, and the orientation of the target obstacle is determined based on multiple feature points at the same height of the target obstacle, and the orientation angle between the target obstacle and the target vehicle is calculated in combination with the orientation of the target vehicle, so that the relative orientation between the target obstacle and the target vehicle at different distances can be calculated according to the feature points at different heights in the image of the target obstacle, so as to be suitable for the recognition of large obstacles in various scenarios, thereby improving the recognition ability of automatic parking.

[0072] In one embodiment, the above-mentioned orientation angle calculation module 34 can be further configured as follows: if there are multiple ground feature points among the feature points, the orientation angle of the target obstacle is calculated based on the multiple ground feature points; wherein the ground feature point is the feature point where the target obstacle contacts the ground.

[0073] In one embodiment, the orientation angle calculation module 34 may be further configured as follows: if the number of the ground feature points is greater than 2, two ground feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

[0074] In one embodiment, the orientation angle calculation module 34 may be further configured as follows: if there are multiple central feature points at the same height among the feature points, the orientation angle of the target obstacle is calculated based on the multiple central feature points; wherein the central feature point is the geometric center point of the target obstacle.

[0075] In one embodiment, the orientation angle calculation module 34 may be further configured as follows: if there are multiple top feature points among the feature points, the orientation angle of the target obstacle is calculated based on the multiple top feature points; wherein the top feature point is the top feature point of the target obstacle.

[0076] In one embodiment, the orientation angle calculation module 34 may be further configured as follows: if the number of upper feature points is greater than 2, two upper feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

[0077] In one embodiment, the orientation angle calculation module 34 may be further configured to: if multiple orientation angles are calculated, remove abnormal values ​​from the multiple orientation angles to obtain a final orientation angle of the target obstacle.

[0078] Below, reference Figure 4 The electronic device according to the embodiment of the present application is described. The electronic device may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signal from them.

[0079] Figure 4 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0080] like Figure 4 As shown, the electronic device 10 includes one or more processors 11 and a memory 12 .

[0081] The processor 11 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0082] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may run the program instructions to implement the methods of the various embodiments of the present application described above and / or other desired functions. Various contents such as input signals, signal components, noise components, etc. may also be stored in the computer-readable storage medium.

[0083] In one example, the electronic device 10 may further include: an input device 13 and an output device 14 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0084] When the electronic device is a stand-alone device, the input device 13 may be a communication network connector, which is used to receive the collected input signals from the first device and the second device.

[0085] In addition, the input device 13 may also include, for example, a keyboard, a mouse, etc.

[0086] The output device 14 can output various information to the outside, including the determined distance information, direction information, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0087] Of course, to simplify, Figure 4 Only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 10 may also include any other appropriate components.

[0088] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.

[0089] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0090] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0091] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0092] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0093] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0094] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0095] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0096] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for identifying the posture of a large obstacle, characterized in that: Applied to a target vehicle, the large obstacle posture recognition method includes: Acquire image data containing target obstacles; Performing region segmentation on the image data to obtain a bounding box of the target obstacle; wherein the bounding box of the target obstacle represents a minimum geometric frame containing the target obstacle; According to the feature information of the target obstacle, matching the feature points of the target obstacle in the bounding box; wherein the feature points represent landmark points at specific positions on the target obstacle; Selecting multiple target feature points at the same height among the feature points, and calculating the orientation angle of the target obstacle; wherein the orientation angle represents the angle between the orientations of the target obstacle and the target vehicle; The selecting a plurality of target feature points at the same height from the feature points and calculating the orientation angle of the target obstacle comprises: If there are multiple central feature points at the same height among the feature points, the orientation angle of the target obstacle is calculated according to the multiple central feature points; wherein the central feature points include the center points of the front and rear axles of the vehicle; Among them, if the image data contains the ground feature point of the target obstacle, the height coordinate of the center feature point in the vehicle coordinate system is calculated based on the ground plane; if the image data does not contain the ground feature point of the target obstacle, the height coordinate of the center feature point in the vehicle coordinate system is calculated based on the feature points of the front and rear wheels.

2. The large obstacle posture recognition method according to claim 1, characterized in that: The selecting a plurality of target feature points at the same height from the feature points and calculating the orientation angle of the target obstacle comprises: If there are multiple ground feature points among the feature points, the orientation angle of the target obstacle is calculated based on the multiple ground feature points; wherein the ground feature point is a feature point where the target obstacle contacts the ground.

3. The large obstacle posture recognition method according to claim 2, characterized in that: If there are multiple ground feature points among the feature points, calculating the orientation angle of the target obstacle according to the multiple ground feature points includes: If the number of the ground feature points is greater than 2, the two ground feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

4. The large obstacle posture recognition method according to claim 1, characterized in that: The selecting a plurality of target feature points at the same height from the feature points and calculating the orientation angle of the target obstacle comprises: If there are multiple top feature points among the feature points, the orientation angle of the target obstacle is calculated according to the multiple top feature points; wherein the top feature point is the top feature point of the target obstacle.

5. The large obstacle posture recognition method according to claim 4, characterized in that: If there are multiple top feature points among the feature points, calculating the orientation angle of the target obstacle according to the multiple top feature points includes: If the number of the top feature points is greater than 2, two top feature points with the farthest relative distance are selected to calculate the orientation angle of the target obstacle.

6. The large obstacle posture recognition method according to claim 1, characterized in that: The selecting a plurality of target feature points at the same height from the feature points and calculating the orientation angle of the target obstacle comprises: If multiple orientation angles are calculated, outliers among the multiple orientation angles are eliminated to obtain a final orientation angle of the target obstacle.

7. A large obstacle posture recognition device, characterized in that: The large obstacle posture recognition device is provided on the target vehicle and includes: An image acquisition module, used to acquire image data containing target obstacles; A boundary segmentation module, used to perform region segmentation on the image data to obtain a boundary box of the target obstacle; wherein the boundary box of the target obstacle represents a minimum geometric frame containing the target obstacle; A feature point matching module, used to match the feature points of the target obstacle within the bounding box according to the feature information of the target obstacle; wherein the feature points represent landmark points at specific locations on the target obstacle; A heading angle calculation module, used to select a plurality of target feature points at the same height among the feature points, and calculate the heading angle of the target obstacle; wherein the heading angle represents the angle between the directions of the target obstacle and the target vehicle; The orientation angle calculation module is further configured as follows: If there are multiple central feature points at the same height among the feature points, the orientation angle of the target obstacle is calculated according to the multiple central feature points; wherein the central feature points include the center points of the front and rear axles of the vehicle; Among them, if the image data contains the ground feature point of the target obstacle, the height coordinate of the center feature point in the vehicle coordinate system is calculated based on the ground plane; if the image data does not contain the ground feature point of the target obstacle, the height coordinate of the center feature point in the vehicle coordinate system is calculated based on the feature points of the front and rear wheels.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the posture recognition method of a large obstacle as described in any one of claims 1 to 6.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is used to execute the large obstacle posture recognition method described in any one of claims 1-6 above.