Vehicle driving track score obtaining method and device and storage medium

Abstract

The invention provides a vehicle driving track score obtaining method and device and a storage medium, and the method comprises the steps: obtaining object data corresponding to a road object according to the camera data of a road test camera and offline map data, the object data comprising dynamic data and static data; matching a target driving track recorded by the target vehicle with the dynamic data to obtain a target vehicle track in the dynamic data; and according to the target vehicle track and the object data, determining a target score corresponding to the target driving track. The vehicle data corresponding to the target vehicle is used for training the automatic driving model, and the target score can provide a judgment basis for training the automatic driving model. Therefore, the vehicle driving track score obtaining method provided by the invention can provide data support for training an automatic driving model.

Description

Technical Field

[0001] This application belongs to the field of big data technology, and particularly relates to a method, device, and storage medium for obtaining vehicle driving trajectory scores. Background Art

[0002] At present, autonomous driving generally adopts a two-stage model. The first-stage model is the perception model of an autonomous driving vehicle; the second-stage model is the decision-making and control model of an autonomous driving vehicle. However, combining the two-stage model into a one-stage model, that is, an end-to-end autonomous driving model, is gradually becoming a trend.

[0003] In addition to the differences in technical routes, the end-to-end autonomous driving model and the two-stage autonomous driving model also have differences in their requirements for training data. In terms of data scale, the two-stage autonomous driving model is a medium model, while the end-to-end autonomous driving model is a large model. The large model has higher requirements for data quality than the medium model.

[0004] Therefore, how to screen higher-quality vehicle data is an urgent problem to be solved. Summary of the Invention

[0005] In view of the problems existing in the above-mentioned prior art, a method, device, and computer-readable storage medium for obtaining vehicle driving trajectory scores are proposed. Using this method, device, and computer-readable storage medium for obtaining vehicle driving trajectory scores, driving data with higher data quality can be screened.

[0006] This application provides the following solutions.

[0007] In a first aspect, this application provides a method for obtaining a vehicle driving trajectory score, including:

[0008] Based on the camera data of a road test camera and offline map data, object data corresponding to road objects is obtained, and the object data includes dynamic data and static data;

[0009] The target driving trajectory recorded by the target vehicle is matched with the dynamic data to obtain the target vehicle track in the dynamic data;

[0010] Based on the target vehicle track and the object data, the target score corresponding to the target driving trajectory is determined, and the vehicle data corresponding to the target vehicle is used to train the autonomous driving model.

[0011] In some possible embodiments, obtaining object data corresponding to road objects based on the camera data of a road test camera and offline map data includes:

[0012] Based on the camera data of a road test camera and offline map data, relevant parameters of the road test camera that collected the camera data are calibrated;

[0013] Detect objects in the camera data to obtain object data corresponding to road objects.

[0014] In some possible embodiments, the relevant parameters of the roadside camera include the internal parameters and external parameters of the roadside camera.

[0015] In some possible embodiments, detecting objects in the camera data to obtain object data corresponding to road objects includes:

[0016] Dynamically identify objects in the camera data to obtain dynamic data, where the dynamic data includes at least one type of data among vehicle dynamic data, pedestrian dynamic data, cyclist dynamic data, and animal dynamic data;

[0017] Staticly identify objects in the camera data to obtain static data, where the static data includes road obstacle data.

[0018] In some possible embodiments, dynamically identifying objects in the camera data to obtain dynamic data includes:

[0019] Obtain dynamic data through single-frame 2D detection and 3D detection, and 2D tracking and 3D tracking based on the camera matrix.

[0020] In some possible embodiments, it further includes:

[0021] Detect the status of traffic lights in the camera data to obtain signal data of the traffic lights;

[0022] Judging the target score corresponding to the target driving trajectory according to the target vehicle track and object data, including:

[0023] Judge the target score corresponding to the target driving trajectory according to the target vehicle track, object data, and signal data.

[0024] In some possible embodiments, the signal data includes signal light position data and traffic light status data;

[0025] Detecting the status of traffic lights in the camera data to obtain signal data of the traffic lights includes:

[0026] Detect the status of traffic lights in the camera data to obtain the traffic light status data of the traffic lights;

[0027] Determine the signal position data of the traffic lights according to the offline map data and / or camera data.

[0028] In some possible embodiments, matching the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data, including:

[0029] Inflate the target driving trajectory recorded by the target vehicle within the target time period into edges to obtain a reference polygon;

[0030] Inflate the vehicle tracks in the dynamic data within the target time period into edges to obtain a plurality of candidate polygons;

[0031] Determine the target polygon with the largest intersection with the reference polygon among the plurality of candidate polygons, and the vehicle track corresponding to the target polygon is the target vehicle track.

[0032] In some possible embodiments, judging the target score corresponding to the target driving trajectory according to the target vehicle track, object data and signal data includes:

[0033] Input the target vehicle track, object data and signal data into a trained artificial intelligence model to obtain the target score corresponding to the target driving trajectory.

[0034] In some possible embodiments, judging the target score corresponding to the target driving trajectory according to the target vehicle track, object data and signal data includes:

[0035] Judge the target score corresponding to the target driving trajectory through a vector network according to the target vehicle track, object data and signal data.

[0036] In some possible embodiments, judging the target score corresponding to the target driving trajectory through a vector network according to the target vehicle track, object data and signal data includes:

[0037] Extract the feature vectors of the objects in the object data and signal data, and the objects are in one-to-one correspondence with the feature vectors;

[0038] Construct a target sequence according to the target vehicle track and extract the target vector of the target sequence;

[0039] Establish high-order interactions between road objects in the object data and signal data to obtain high-order features of the target sequence;

[0040] Input the high-order features of the target sequence into a fitter to obtain the target score corresponding to the target driving trajectory.

[0041] In a second aspect, the present application provides an apparatus for obtaining a vehicle driving trajectory score, including:

[0042] A data acquisition module, configured to obtain object data corresponding to road objects according to the camera data of a road test camera and offline map data, where the object data includes dynamic data and static data;

[0043] A matching module, configured to match a target driving trajectory recorded by a target vehicle with dynamic data to obtain a target vehicle track in the dynamic data;

[0044] A judging module, configured to judge a target score corresponding to the target driving trajectory according to the target vehicle track and object data, and vehicle data corresponding to the target vehicle is used to train an autonomous driving model.

[0045] In a third aspect, the present application provides a device for obtaining a vehicle driving trajectory score, including:

[0046] At least one processor; and, a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute: the above-mentioned method for obtaining a vehicle driving trajectory score.

[0047] In a fourth aspect, the present application provides a computer-readable storage medium, and the computer-readable storage medium stores a program, which when executed by a multi-core processor, enables the multi-core processor to execute the above-mentioned method for obtaining a vehicle driving trajectory score.

[0048] The present application provides a method, device and storage medium for obtaining a vehicle driving trajectory score. The method includes: obtaining object data corresponding to road objects according to camera data of a road test camera and offline map data, where the object data includes dynamic data and static data; matching a target driving trajectory recorded by a target vehicle with the dynamic data to obtain a target vehicle track in the dynamic data; judging a target score corresponding to the target driving trajectory according to the target vehicle track and the object data. Vehicle data corresponding to the target vehicle is used to train an autonomous driving model, and the target score can provide a judgment basis for training the autonomous driving model. Therefore, the method for obtaining a vehicle driving trajectory score provided by the present application can provide data support for training an autonomous driving model.

[0049] Other advantages of the present application will be explained in more detail in conjunction with the following description and drawings.

[0050] It should be understood that the above description is only an overview of the technical solution of the present application, so as to be able to understand the technical means of the present application more clearly, and thus can be implemented according to the content of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically illustrates the specific embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] By reading the detailed description of the exemplary embodiments below, those of ordinary skill in the art will understand the advantages and benefits described herein as well as other advantages and benefits. The drawings are only for the purpose of illustrating the exemplary embodiments and are not considered to be a limitation of the present application. In the drawings:

[0052] Figure 1 It is a schematic structural diagram of a hardware operating environment provided by an embodiment of the present application;

[0053] Figure 2 It is a schematic flowchart of a method for obtaining a vehicle driving trajectory score provided by an embodiment of the present application;

[0054] Figure 3 It is a schematic flowchart of a method for determining a target driving trajectory provided by an embodiment of the present application;

[0055] Figure 4 It is a schematic diagram of a device for obtaining a vehicle driving trajectory score provided by an embodiment of the present application;

[0056] In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Embodiments

[0057] The exemplary embodiments of the present application will be described in more detail below with reference to the drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be fully conveyed to those skilled in the art.

[0058] In the description of the embodiments of the present application, it should be understood that terms such as "including" or "having" are intended to indicate the presence of the disclosed features, numbers, steps, actions, components, parts, or combinations thereof in the present specification, and do not exclude the possibility of the presence of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0059] Unless otherwise specified, " / " means "or". For example, A / B may mean A or B; herein, "and / or" is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B may mean: A exists alone, A and B exist simultaneously, and B exists alone.

[0060] Terms such as "first" and "second" are only used for convenience of description to distinguish the same or similar technical features, and cannot be understood as indicating or implying the relative importance or quantity of these technical features. Thus, the features defined by "first", "second", etc. may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of the term "plurality" is two or more than two.

[0061] In addition, it should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will describe this application in detail with reference to the drawings and in combination with the embodiments.

[0062] As Figure 1 shown, Figure 1 is a schematic structural diagram of a hardware operating environment provided by an embodiment of this application.

[0063] It should be noted that, Figure 1 it can be a schematic structural diagram of the hardware operating environment of the device for obtaining the vehicle driving trajectory score. Based on the device for obtaining the vehicle driving trajectory score in the embodiments of this application, it can be a terminal device such as a PC or a portable computer.

[0064] As Figure 1 shown, the device for obtaining the vehicle driving trajectory score may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard). The user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may optionally be a storage device independent of the aforementioned processor 1001.

[0065] Those skilled in the art can understand that, Figure 1 the structure of the device for obtaining the vehicle driving trajectory score shown in

[0066] Figure 2 does not constitute a limitation on the device for obtaining the vehicle driving trajectory score, and it may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0067] As Figure 2 shown, the method for obtaining the vehicle driving trajectory score provided by the embodiments of this application may include:

[0068] S201: Obtain object data corresponding to road objects based on the captured data of the roadside camera and the offline map data, where the object data includes dynamic data and static data.

[0069] In the embodiments of the present application, the roadside camera may include road monitoring cameras or other cameras installed on the roadside, which are not limited in the embodiments of the present application. The captured data of the roadside camera generally refers to the captured data collected by the roadside camera.

[0070] The offline map data in the embodiments of the present application may be high-precision map data. A high-precision map is a map with high measurement accuracy and high element richness, which includes specific map elements such as lane lines, lane line endpoints, lanes, road edges, road centerlines, ground symbols, ground symbol corner points, barriers, curbs, lamp posts, road signs, fences, bridges and tunnels, traffic lights, etc. Among them, traffic lights, in addition to position information, also have topological information with lanes. The road objects in the embodiments of the present application include roads, vehicles, pedestrians, and other objects that can affect vehicle driving, which are not limited in the embodiments of the present application.

[0071] In the embodiments of the present application, the relevant parameters of the roadside camera that captures the captured data can be calibrated according to the captured data of the roadside camera and the offline map data, and then the objects in the captured data can be detected to obtain the object data corresponding to the road objects. The relevant parameters of the roadside camera may include the internal parameters and external parameters of the roadside camera. Among them, the internal parameters of the roadside camera are used to describe the inherent characteristics of the camera itself and are independent of the position and orientation of the camera. The internal parameters may include one or more of focal length, principal point, pixel size, and distortion coefficient. Among them, the external parameters of the roadside camera are used to describe the position and orientation of the camera relative to the world coordinate system, and the external parameters may include a rotation matrix and / or a translation vector.

[0072] Specifically, the calibration of the external parameters of the camera can mainly rely on the acquisition of control points. In the actual operation process, when the internal parameters of the roadside camera are known, the feature point p1 on the image can be obtained within the area covered by the roadside camera, and then the world coordinate P1 of the feature point p1 can be calculated through the high-precision map to form a pair of control points. After having multiple groups of control points, the external parameters of the imaging device can be solved through the PnP algorithm. When the internal parameters of the roadside camera are not obtained, first obtain multiple groups of control points {(p1, P1), (p2, P2),..., (pn, Pn)}, and then optimize the internal parameters and external parameters of the roadside camera based on the reprojection error and the least squares algorithm.

[0073] In practical applications, the embodiments of the present application can dynamically identify objects in the camera data to obtain dynamic data. The dynamic data may include vehicle dynamic data, pedestrian dynamic data, cyclist dynamic data, animal dynamic data, etc. Then, the objects in the camera data are statically identified to obtain static data, which may include road obstacle data. The road obstacle data may include data of various different obstacles on the road, and the embodiments of the present application do not make limitations here.

[0074] The embodiments of the present application can obtain dynamic data through the detection of dynamic objects. Specifically, the present application can obtain dynamic data through single-frame 2D detection and 3D detection, and 2D tracking and 3D tracking based on the camera matrix. The purpose of 2D detection and tracking is to improve the accuracy of 3D detection and tracking. The 2D detection outputs the position and size of each object, that is: (x, y, w, l). The 3D detection outputs the position, size, direction and speed of each object, that is: (x, y, z, l, w, h, yaw, vx, vy). 2D or 3D tracking is to determine the motion trajectory of the same object at different times (or different frames). A 3D tracking trajectory can be expressed as {(T1, O1), (T2, O2),..., (Tn, On)}. The 3D tracking of the objects in the embodiments of the present application may include tracking with the same camera and tracking between different cameras, and the embodiments of the present application do not make limitations here.

[0075] In addition to obtaining the camera data of the roadside camera and the offline map data, the embodiments of the present application can also detect the status of traffic lights in the camera data to obtain signal data of the traffic lights. As a possible implementation manner, in the embodiments of the present application, the signal data includes signal light position data and traffic light status data. The embodiments of the present application can detect the status of traffic lights in the camera data to obtain the signal light status data of the traffic lights; and determine the signal position data of the traffic lights according to the offline map data and / or the camera data. As an example, the embodiments of the present application can update the relevant data of the traffic lights in the offline map data according to the status data of the traffic lights to obtain the updated offline map data, so as to score the driving trajectory more accurately subsequently.

[0076] S202: Match the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data.

[0077] In the embodiments of the present application, the key to extracting the driving trajectory of the target vehicle is to match the driving trajectory recorded by the target vehicle with the driving trajectories of the dynamic vehicles in the dynamic data obtained by the roadside camera. The matched driving trajectory is the driving trajectory of the target vehicle. Specifically, the target driving trajectory recorded by the target vehicle within the target time period can be inflated into edges to obtain a reference polygon. Then, the vehicle tracks in the dynamic data within the target time period are inflated into edges to obtain a plurality of candidate polygons; and the target polygon with the largest intersection with the reference polygon among the plurality of candidate polygons is determined, and the vehicle track corresponding to the target polygon is the target vehicle track.

[0078] As a possible implementation manner, the embodiments of the present application can match the traffic lights detected in real time with the traffic lights in the high-precision map to obtain the status sequence of the traffic lights in the high-precision map, providing a basis for the driving actions of the vehicle. When matching the traffic lights, first, according to the camera internal parameters, position, and the contour information of the traffic lights, a cone is generated. Combining the high-precision map information, the traffic light closest to the cone is regarded as the matched traffic light.

[0079] S203: According to the target vehicle track and object data, determine the target score corresponding to the target driving trajectory, and the vehicle data corresponding to the target vehicle is used to train the autonomous driving model.

[0080] After obtaining the signal data of the traffic lights, the embodiments of the present application can determine the target score corresponding to the target driving trajectory according to the target vehicle track, object data, and signal data. As a possible implementation manner, an expert can determine the target score corresponding to the target driving trajectory according to the target vehicle track, signal data, and object data. As another possible implementation manner, the embodiments of the present application can input the target vehicle track, signal data, and object data into a trained artificial intelligence model to obtain the target score corresponding to the target driving trajectory.

[0081] As another possible implementation manner, the embodiments of the present application can also determine the target score corresponding to the target driving trajectory through a VectorNet according to the target vehicle track, signal data, and object data. Specifically, as Figure 3 shown, the steps for the embodiments of the present application to determine the target score corresponding to the target driving trajectory through a VectorNet can include:

[0082] S301: Extract the feature vectors of the objects in the object data and signal data, and the objects are in one-to-one correspondence with the feature vectors.

[0083] Among them, the high-precision map elements and the trajectories of dynamic objects can both be represented as vector sequences. The embodiments of the present application can obtain the features of the vector sequences based on the local graph network.

[0084] S302: Construct a target sequence based on the target vehicle's trajectory and extract the target vector of the target sequence.

[0085] As an example, the connection from the current position to the next position of the target vehicle can form the target sequence. The construction features of the target sequence can include: starting position, ending position, starting direction, ending direction, and start-end distance. Based on the fully connected network, the embodiments of the present application can extract the sequence features. Then, the target vector can be formed according to the sequence features of the target sequence.

[0086] S303: Establish high-order interactions between the road objects in the object data and the signal data to obtain the high-order features of the target sequence.

[0087] As a possible implementation, the embodiments of the present application can establish high-order interactions between each object (including the target sequence) based on the self-attention mechanism to obtain the high-order features of each object and the high-order features of the target sequence.

[0088] S304: Input the high-order features of the target sequence into the fitter to obtain the target score corresponding to the target driving trajectory.

[0089] In summary, for the method for obtaining the vehicle driving trajectory score provided by the embodiments of the present application, in addition to using the vehicle's own motion trajectory, object data is also used. The object data includes the motion trajectories of surrounding dynamic objects, static environment information, and traffic light status information to judge the target score of the target driving trajectory recorded by the target vehicle. The present application also judges the target score corresponding to the target driving trajectory through the VectorNet. The vehicle data corresponding to the target vehicle is used to train the autonomous driving model, and the target score can provide a judgment basis for training the autonomous driving model. In this way, the method for obtaining the vehicle driving trajectory score provided by the embodiments of the present application can provide data support for training the autonomous driving model.

[0090] In the description of this specification, the descriptions made with reference to terms such as "some possible implementations", "some implementations", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the implementation or example are included in at least one implementation or example of the present application, and the above terms do not necessarily refer to the same implementation or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more implementations or examples. In addition, without contradiction, those skilled in the art can combine and combine the different implementations or examples described in this specification and the features of different implementations or examples.

[0091] Regarding the method flowchart of the embodiments of the present application, certain operations are described as different steps executed in a certain order. Such a flowchart is illustrative rather than restrictive. Certain steps described herein can be grouped together and executed in a single operation, or certain steps can be split into multiple sub-steps, and certain steps can be executed in an order different from that shown herein. Each step shown in the flowchart can be implemented in any manner by any circuit structure and / or tangible mechanism (e.g., by software running on a computer device, hardware (e.g., a processor or chip-implemented logic function), etc., and / or any combination thereof).

[0092] Those skilled in the art can understand that in the methods described in the above specific embodiments, the written order of each step does not mean a strict execution order, and the specific execution order of each step should be determined according to its function and possible internal logic.

[0093] According to the method for obtaining the vehicle driving trajectory score provided in the above embodiments, the embodiments of the present application also provide an apparatus for obtaining the vehicle driving trajectory score.

[0094] As Figure 4 shown, the apparatus for obtaining the vehicle driving trajectory score provided by the embodiments of the present application includes:

[0095] A data acquisition module 100, configured to obtain object data corresponding to road objects according to the camera data of the road test camera and the offline map data, where the object data includes dynamic data and static data;

[0096] A matching module 200, configured to match the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data;

[0097] A judgment module 300, configured to judge the target score corresponding to the target driving trajectory according to the target vehicle track and the object data, and the vehicle data corresponding to the target vehicle is used to train the autonomous driving model.

[0098] It should be noted that the apparatus for obtaining the vehicle driving trajectory score in the embodiments of the present application can implement each process of the embodiments of the foregoing method for obtaining the vehicle driving trajectory score, and achieve the same effects and functions, which will not be elaborated here.

[0099] According to some embodiments of the present application, there is provided an apparatus for obtaining a vehicle driving trajectory score according to an embodiment of the present application, for executing Figure 2The method for obtaining the vehicle driving trajectory score shown, the device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described in the above embodiments.

[0100] According to some embodiments of the present application, there is provided a non-volatile computer storage medium for the method of obtaining a vehicle driving trajectory score, on which computer-executable instructions are stored, and the computer-executable instructions are configured to be executed when run by a processor: the method described in the above embodiments.

[0101] Computer-readable media includes both permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory, read-only memory, electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. Additionally, although the operations of the method of the present application are described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Further, certain steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple sub-steps for execution.

[0102] Although the spirit and principles of the present application have been described above with reference to several specific embodiments, it should be understood that the present application is not limited to the disclosed specific embodiments, and the division of each aspect does not mean that the features in these aspects cannot be combined. The present application aims to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for obtaining a vehicle driving trajectory score, characterized in that: include: Obtain object data corresponding to road objects according to the camera data of the road test camera and the offline map data, wherein the object data includes dynamic data and static data; Matching the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data; The target score corresponding to the target driving trajectory is determined based on the target vehicle trajectory and object data, and the vehicle data corresponding to the target vehicle is used to train the autonomous driving model.

2. The method according to claim 1, characterized in that The method of obtaining object data corresponding to road objects according to the camera data of the road test camera and the offline map data includes: Calibrate relevant parameters of the road test camera that collects the camera data according to the camera data of the road test camera and the offline map data; The objects in the camera data are detected to obtain object data corresponding to the road objects.

3. The method according to claim 2, characterized in that The relevant parameters of the road test camera include internal parameters and external parameters of the road test camera.

4. The method according to claim 2, characterized in that: The detecting the object in the camera data to obtain object data corresponding to the road object includes: Dynamically identifying the objects in the camera data to obtain dynamic data, wherein the dynamic data includes at least one type of data selected from vehicle dynamic data, pedestrian dynamic data, cyclist dynamic data, and animal dynamic data; Static recognition is performed on the objects in the camera data to obtain static data, wherein the static data includes road obstacle data.

5. The method according to claim 4, characterized in that The step of dynamically identifying the object in the camera data to obtain dynamic data includes: Dynamic data is obtained through single-frame 2D detection and 3D detection, and 2D tracking and 3D tracking based on the camera matrix.

6. The method according to claim 1, characterized in that Also includes: Detecting the state of the traffic light in the camera data to obtain signal data of the traffic light; The step of determining a target score corresponding to the target driving trajectory according to the target vehicle trajectory and the object data includes: The target score corresponding to the target driving trajectory is determined based on the target vehicle trajectory, the object data and the signal data.

7. The method according to claim 6, characterized in that The signal data includes signal light position data and traffic light status data; Detecting the state of the traffic light in the camera data to obtain signal data of the traffic light includes: Detecting the state of the traffic light in the camera data to obtain the traffic light state data of the traffic light; The signal position data of the traffic light is determined according to the offline map data and / or the camera data.

8. The method according to claim 1, characterized in that The step of matching the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data includes: Expand the target driving trajectory recorded by the target vehicle in the target time period into edges to obtain a reference polygon; Expanding the vehicle tracks in the dynamic data within the target time period into edges to obtain a plurality of polygons to be selected; A target polygon having the largest intersection with the reference polygon among the multiple polygons to be selected is determined, and the vehicle track corresponding to the target polygon is the target vehicle track.

9. The method according to claim 6, characterized in that The step of determining a target score corresponding to the target driving trajectory according to the target vehicle trajectory, the object data, and the signal data includes: The target vehicle track, the object data and the signal data are input into a trained artificial intelligence model to obtain a target score corresponding to the target driving track.

10. The method according to claim 6, characterized in that The step of determining a target score corresponding to the target driving trajectory according to the target vehicle trajectory, the object data, and the signal data includes: According to the target vehicle track, the object data and the signal data, a target score corresponding to the target driving track is determined through a vector network.

11. The method according to claim 10, characterized in that The step of determining a target score corresponding to the target driving trajectory through a vector network according to the target vehicle trajectory, the object data, and the signal data includes: Extracting feature vectors of objects in the object data and the signal data, wherein the objects correspond one-to-one to the feature vectors; Constructing a target sequence according to the target vehicle track, and extracting a target vector of the target sequence; Establishing high-order interactions between the object data and road objects in the signal data to obtain high-order features of the target sequence; The high-order features of the target sequence are input into a fitter to obtain a target score corresponding to the target driving trajectory.

12. A device for obtaining a vehicle driving trajectory score, characterized in that: include: A data acquisition module, used to obtain object data corresponding to road objects according to the camera data of the road test camera and the offline map data, wherein the object data includes dynamic data and static data; A matching module, used for matching the target driving trajectory recorded by the target vehicle with the dynamic data to obtain the target vehicle track in the dynamic data; A judgment module is used to judge the target score corresponding to the target driving trajectory according to the target vehicle trajectory and the object data, and the vehicle data corresponding to the target vehicle is used to train the automatic driving model.

13. A device for obtaining a vehicle driving trajectory score, characterized in that: include: at least one processor; And, a memory communicatively connected to at least one processor; wherein the memory stores instructions executable by at least one processor, and the instructions are executed by at least one processor to enable at least one processor to execute: a method as described in any one of claims 1-11.

14. A computer-readable storage medium storing a program, wherein when the program is executed by a multi-core processor, the multi-core processor executes the method according to any one of claims 1 to 11.

Images