Positioning accuracy test method, equipment, storage medium and program product

Abstract

The embodiment of the present application provides a positioning accuracy test method, device, storage medium and program product, the method includes determining, for each feature point of a semantic element, the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point, determining the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point, and determining the error result corresponding to the semantic element based on the target true value distance and the target measurement distance, and then determining the relative positioning accuracy of the positioning trajectory to be processed based on the error results corresponding to the multiple semantic elements. The positioning accuracy test method provided in the embodiment of the present application determines the error result of the distance between the vehicle and the semantic elements around the driving trajectory by using the laser point cloud as the true value reference data, and then obtains the relative positioning accuracy of the vehicle positioning trajectory.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of autonomous driving technology, and in particular to a positioning accuracy testing method, device, storage medium, and program product. Background Art

[0002] The autonomous driving system is a complex system that integrates multiple technologies such as perception, positioning, planning and control. High-precision positioning technology is the basis for the stable, accurate and fast operation of the entire system. Therefore, positioning accuracy is a key indicator for evaluating the entire system.

[0003] In the related art, a navigation or post-processing trajectory result with higher accuracy is usually used as the true value of positioning, such as an IE trajectory or a real-time RTK trajectory, and then the positioning accuracy is determined based on the true value of positioning and the result of real-time positioning.

[0004] However, in the process of implementing this application, the inventors found that there are at least the following problems in the prior art: the positioning accuracy obtained by the above method is the global positioning accuracy, that is, the absolute positioning accuracy. In the current autonomous driving solution, compared with the absolute positioning accuracy, the system pays more attention to the distance between the vehicle and some scene markers under the local map, that is, the relative positioning accuracy. How to accurately evaluate the relative positioning accuracy is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The embodiments of the present application provide a positioning accuracy testing method, device, storage medium and program product to achieve relative positioning accuracy testing.

[0006] In a first aspect, an embodiment of the present application provides a positioning accuracy testing method, comprising:

[0007] Acquire a target point cloud frame corresponding to a target feature point, and first points of the same name corresponding to a plurality of the feature points in the target point cloud frame; the plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to the positioning trajectory to be processed;

[0008] Acquire a target image frame corresponding to a plurality of the feature points, and second points of the same name corresponding to the plurality of the feature points in the target image frame, wherein the target image frame is collected during the driving process of the target vehicle based on the positioning trajectory to be processed;

[0009] Acquire target trajectory nodes respectively corresponding to a plurality of the feature points in the positioning trajectory to be processed;

[0010] For each feature point of a semantic element, determining a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determining a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determining an error result corresponding to the semantic element based on the target true value distance and the target measured distance;

[0011] The relative positioning accuracy of the positioning trajectory to be processed is determined according to the error results corresponding to the plurality of semantic elements.

[0012] In a possible design, obtaining a target point cloud frame corresponding to the target feature point includes:

[0013] Obtaining a positioning trajectory to be processed, a vector semantic map corresponding to the positioning trajectory to be processed, and a corresponding laser point cloud frame;

[0014] Selecting feature points corresponding to a plurality of semantic elements respectively from the vector semantic map; the plurality of semantic elements are located within a preset range of the positioning trajectory to be processed;

[0015] A plurality of target point cloud frames corresponding to the feature points are selected from the laser point cloud frames.

[0016] In a possible design, the selecting a plurality of target point cloud frames corresponding to the feature points from the laser point cloud frames includes:

[0017] Acquire sensor data corresponding to the positioning trajectory to be processed;

[0018] Based on the positioning trajectory to be processed and the sensor data, the coordinates of each point cloud in the laser point cloud frame are converted to the same coordinate system to obtain a corrected laser point cloud frame;

[0019] A plurality of target point cloud frames corresponding to the feature points are selected from the corrected laser point cloud frames.

[0020] In a possible design, determining the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point includes:

[0021] Obtaining a distance measurement result corresponding to a second point of the same name corresponding to the feature point; the distance measurement result is a distance between the second point of the same name generated in the process of generating the positioning trajectory to be processed and the target trajectory node;

[0022] The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes:

[0023] The difference between the distance measurement result and the target true value distance is determined as the distance measurement error corresponding to the semantic element.

[0024] In a possible design, determining the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point includes:

[0025] Determine the true value coordinates of the target trajectory node corresponding to the feature point;

[0026] The coordinates of the feature points and the true value coordinates of the corresponding target trajectory nodes are converted into a target coordinate system, and the target relative distance between the feature points of the semantic elements and the target trajectory nodes is calculated;

[0027] The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes:

[0028] The difference between the target true value distance and the target relative distance is determined as the relative positioning error corresponding to the semantic element.

[0029] In a possible design, determining the true value coordinates of the target trajectory node includes:

[0030] Obtaining a true value trajectory corresponding to the positioning trajectory to be processed;

[0031] A neighboring node adjacent to the target trajectory node is searched from the true trajectory, and based on an interpolation algorithm, the true coordinates of the target trajectory node are determined according to the coordinates and timestamps of the neighboring nodes and the timestamp of the target trajectory node.

[0032] In a possible design, determining the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point includes:

[0033] If the semantic element is an aerial semantic element, the target true value distance is determined according to the coordinates of the first point of the same name corresponding to the feature point.

[0034] In a possible design, determining the target true value distance between the target vehicle and the aerial semantic element according to the coordinates of the first same-name point corresponding to the feature point of the aerial semantic element includes:

[0035] If the aerial semantic element is a planar semantic element, and the plane where the aerial semantic element is located is perpendicular to the driving direction of the target vehicle, the longitudinal coordinate component in the coordinates of the first same-name point corresponding to the feature point of the aerial semantic element is determined as the longitudinal true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point;

[0036] The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes:

[0037] Based on the longitudinal true value distance and the target measurement distance, a longitudinal error corresponding to the semantic element is determined.

[0038] In a possible design, determining the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point includes:

[0039] If the semantic element is a ground semantic element, determining a ground plane equation corresponding to the ground semantic element;

[0040] According to the ground plane equation, the second same-name point corresponding to the feature point of the ground semantic element is converted from a two-dimensional feature point to a three-dimensional feature point, and according to the coordinates of the three-dimensional feature point, the target true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point is determined.

[0041] In a possible design, determining the target true value distance between the target vehicle and the ground semantic element according to the coordinates of the three-dimensional feature point includes:

[0042] If the ground plane semantic element is a linear semantic element, and the ground plane semantic element is parallel to the driving direction of the target vehicle, then determining the lateral true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point according to the lateral coordinate of the coordinate of the three-dimensional feature point;

[0043] The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes:

[0044] Based on the lateral true value distance and the target measurement distance, a lateral error corresponding to the semantic element is determined.

[0045] In a possible design, determining the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the plurality of semantic elements includes:

[0046] Based on different scenarios, the positioning trajectory to be processed is divided into multiple segments to obtain multiple sub-trajectories;

[0047] For each sub-trajectory, an error result corresponding to the sub-trajectory is determined, and a relative positioning accuracy of the sub-trajectory is determined according to the error result.

[0048] In a second aspect, an embodiment of the present application provides a positioning accuracy testing device, including:

[0049] An acquisition module is used to acquire a target point cloud frame corresponding to a plurality of feature points, and first points of the same name corresponding to the plurality of feature points in the target point cloud frame; the plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to the positioning trajectory to be processed;

[0050] The acquisition module is further used to acquire target image frames corresponding to the plurality of feature points, and second points of the same name corresponding to the plurality of feature points in the target image frames, wherein the target image frames are collected during the driving process of the target vehicle based on the positioning trajectory to be processed;

[0051] The acquisition module is further used to acquire target trajectory nodes respectively corresponding to the plurality of feature points in the positioning trajectory to be processed;

[0052] a processing module, configured to determine, for each feature point of a semantic element, a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determine a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determine an error result corresponding to the semantic element based on the target true value distance and the target measured distance;

[0053] The processing module is further used to determine the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the multiple semantic elements.

[0054] In a third aspect, an embodiment of the present application provides a positioning accuracy testing device, including: at least one processor and a memory;

[0055] The memory stores computer-executable instructions;

[0056] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the method described in the first aspect and various possible designs of the first aspect.

[0057] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the method described in the first aspect and various possible designs of the first aspect are implemented.

[0058] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the method described in the first aspect and various possible designs of the first aspect.

[0059] The positioning accuracy test method, device, storage medium and program product provided in this embodiment first determine, for each feature point of a semantic element, the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point, determine the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point, and determine the error result corresponding to the semantic element based on the target true value distance and the target measurement distance, and then determine the relative positioning accuracy of the positioning trajectory to be processed based on the error results corresponding to the multiple semantic elements. The positioning accuracy test method provided in the embodiment of the present application determines the error result of the distance between the vehicle and the semantic elements around the driving trajectory by using the laser point cloud as the true value reference data, and then obtains the relative positioning accuracy of the vehicle positioning trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0061] Figure 1 A schematic diagram of the principle of the positioning accuracy testing method provided in the embodiment of the present application;

[0062] Figure 2 A flowchart of a positioning accuracy testing method provided in an embodiment of the present application;

[0063] Figure 3 A schematic diagram of a true value trajectory and a positioning trajectory to be processed provided in an embodiment of the present application;

[0064] Figure 4 A schematic diagram of the structure of a positioning accuracy test device provided in an embodiment of the present application;

[0065] Figure 5 A schematic diagram of the hardware structure of the positioning accuracy testing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0067] Autonomous driving is a complex system that integrates multiple technologies such as perception, positioning, planning and control. Among them, high-precision positioning technology is the basis for the stable, accurate and fast operation of the entire system. At present, the mainstream positioning technologies are based on high-precision map positioning, including high-precision point cloud map positioning and high-precision semantic map positioning. Point cloud maps can provide accurate three-dimensional environmental information for positioning, and semantic maps provide lightweight semantic vector information. Based on the positioning of the above two types of maps, combined with GPS, inertial navigation and other information, high-frequency, high-precision global positioning information and local motion trajectories are output for the downstream environmental perception, planning and control modules of autonomous driving. Environmental perception relies on positioning to build a world model, and planning performs global trajectory planning based on the position of the vehicle output by positioning, and provides vehicle status information for control. Therefore, the positioning module is the premise of the entire autonomous driving system, and its operation accuracy is also a key indicator considered by the entire system.

[0068] In related technologies, higher-precision navigation or post-processing trajectory results are usually used as the true value of positioning, such as real-time differential positioning (Real-Time Kinematic, RTK) trajectory. This type of method uses the IE or RTK trajectory at the same time as the positioning position true value and attitude true value, and projects the results of real-time positioning to the horizontal and vertical directions of the true value trajectory as the positioning position error and heading angle error. However, the positioning accuracy obtained by the above method is the global positioning accuracy, that is, the absolute positioning accuracy. In the current autonomous driving solution, compared with the absolute positioning accuracy, the system pays more attention to the accuracy of the distance between the vehicle and some scene markers under the local map, that is, the relative positioning accuracy of the autonomous driving vehicle in the map topology network. How to accurately evaluate the relative positioning accuracy is a technical problem that needs to be solved urgently.

[0069] In order to solve the above technical problems, the inventors of the present application have found that it is possible to obtain laser point clouds while the vehicle is driving, and use the laser point clouds as true value reference data to determine the error results of the distance between the vehicle and the semantic elements around the driving trajectory, and obtain the relative positioning accuracy of the vehicle positioning trajectory. Based on this, the embodiment of the present application provides a positioning accuracy test method, which uses the laser point cloud as the true value to determine the error results of the distance between the vehicle and the semantic elements around the driving trajectory, and then obtains the relative positioning accuracy of the vehicle positioning trajectory.

[0070] Figure 1 The schematic diagram of the principle of the positioning accuracy test method provided in the embodiment of the present application. Figure 1As shown, the vehicle travels along the road and obtains multiple track nodes of the positioning track to be processed. When traveling to track node 103, the position coordinates of the diagonal filled arrow 102 can be obtained through positioning, while the actual position of the position arrow of the vertical filled arrow 101, the distance between the diagonal filled arrow 102 and the vertical filled arrow 101 does not completely overlap, indicating that there is an error in the distance between the vehicle and the surrounding semantic elements (arrows), and the size of the error can indicate the relative positioning accuracy.

[0071] In the specific implementation process, laser point cloud can be obtained by laser radar during vehicle driving, the true value distance can be determined according to the laser point cloud with higher accuracy, and the error of the measured distance in the positioning process can be determined, so as to obtain the relative positioning accuracy. Taking the tip of the arrow 101 filled with vertical lines as an example, the target feature point corresponding to the arrow tip in the vector semantic map is obtained, the first same-name point in the target point cloud frame corresponding to the target feature point in the laser point cloud is obtained, the second same-name point in the target image frame corresponding to the target feature point is obtained, and the target trajectory node (trajectory node 103) corresponding to the target feature point is obtained. The distance between the first same-name point and the trajectory node 103 is the true value distance L2, and the distance between the second same-name point and the trajectory node 103 is the measured distance L1. According to L1 and L2, the error distance ERR can be obtained. Based on this, the relative positioning accuracy of the positioning trajectory to be processed can be determined based on the feature points of multiple semantic elements (ground arrows, lane lines, dotted lines, aerial signs, etc.) around the entire positioning trajectory to be processed. The positioning accuracy test method provided in this embodiment uses the laser point cloud as the true reference data to determine the error result of the distance between the vehicle and the semantic elements around the driving trajectory, and then obtains the relative positioning accuracy of the vehicle positioning trajectory.

[0072] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0073] Figure 2 This is a flow chart of the positioning accuracy test method provided in the embodiment of the present application. Figure 2 As shown, the method includes:

[0074] 201. Obtain a target point cloud frame corresponding to a target feature point, and first points of the same name corresponding to a plurality of the feature points in the target point cloud frame; the plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to a positioning trajectory to be processed.

[0075] 202. Acquire a target image frame corresponding to a plurality of the feature points, and second points of the same name corresponding to the plurality of the feature points in the target image frame, wherein the target image frame is collected during the driving process of the target vehicle based on the positioning trajectory to be processed.

[0076] 203. Obtain target trajectory nodes corresponding to the plurality of feature points in the positioning trajectory to be processed.

[0077] In this embodiment, the execution subject can be a computer, a tablet, a mobile phone, a server, or other device with computing capabilities, a device in the device, or a chip with the same function.

[0078] Optionally, obtaining the target point cloud frame corresponding to the target feature point may include: obtaining a positioning trajectory to be processed, a vector semantic map corresponding to the positioning trajectory to be processed, and a corresponding laser point cloud frame; selecting feature points corresponding to multiple semantic elements from the vector semantic map; multiple semantic elements are located within a preset range of the positioning trajectory to be processed; and selecting target point cloud frames corresponding to multiple feature points from the laser point cloud frame.

[0079] In some embodiments, in order to improve the accuracy of the data, the laser point cloud can be corrected, and the selecting of multiple target point cloud frames corresponding to the feature points from the laser point cloud frame can include: obtaining the sensor data corresponding to the positioning trajectory to be processed; based on the positioning trajectory to be processed and the sensor data, converting the coordinates of each point cloud in the laser point cloud frame to the same coordinate system to obtain a corrected laser point cloud frame; and selecting multiple target point cloud frames corresponding to the feature points from the corrected laser point cloud frame.

[0080] In this embodiment, the vector semantic map may include the absolute position coordinates (eg, world coordinates under the WGS-84 coordinate system (World Geodetic System—1984 Coordinate System)) of each semantic element within a preset range of the entire positioning trajectory to be processed.

[0081] In this embodiment, the target point cloud frame, the target image frame, and the target feature point can be stored in the truth library by establishing a truth library. Specifically, in the vector semantic map, along the positioning trajectory to be processed, multiple semantic elements (such as arrows, dotted boxes, and aerial signs) are selected within the preset range of the trajectory, and the feature points of each semantic element are determined. The corresponding target image frame and target point cloud frame are selected for each semantic element, and the first homonymous point of the feature point in the target point cloud frame and the second homonymous point in the target image frame are determined. It is also possible to obtain a group of laser points corresponding to the ground around the ground semantic element, so as to subsequently determine the ground plane equation, and then determine the true value distance corresponding to the first homonymous point based on the ground plane equation. The target image frame can be selected from the original image data collected during the vehicle positioning process, that is, during the generation of the positioning trajectory to be processed.

[0082] 204. For each feature point of a semantic element, determine a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determine a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determine an error result corresponding to the semantic element based on the target true value distance and the target measured distance.

[0083] In this embodiment, the error result may include multiple errors, such as ranging error, and relative positioning error of a feature point in a vector semantic map relative to the first same-name point in a target point cloud frame.

[0084] In some embodiments, the calculation method of the ranging error can be: determining the target measurement distance between the target trajectory node corresponding to the feature point and the second homonymous point corresponding to the feature point can include: obtaining the ranging result corresponding to the second homonymous point; the ranging result is the distance between the second homonymous point and the target trajectory node generated in the process of generating the positioning trajectory to be processed; determining the error result corresponding to the semantic element based on the target true value distance and the target measurement distance can include: determining the difference between the ranging result and the target true value distance as the ranging error corresponding to the semantic element.

[0085] Specifically, during the positioning process, that is, during the process of generating the positioning trajectory to be processed, the distance measurement result between the second same-name point and the target feature point can be generated and stored so that it can be directly obtained when performing relative positioning accuracy evaluation.

[0086] In some embodiments, the relative positioning error may be calculated as follows: the determination of the target measured distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point may include: determining the true coordinates of the target trajectory node; converting the coordinates of the feature point and the true coordinates of the corresponding target trajectory node to the target coordinate system, and calculating the target relative distance between the feature point of the semantic element and the target trajectory node; the determination of the error result corresponding to the semantic element based on the target true distance and the target measured distance may include: determining the difference between the target true distance and the target relative distance as the relative positioning error corresponding to the semantic element. Optionally, the determination of the true coordinates of the target trajectory node may include: obtaining the true trajectory corresponding to the positioning trajectory to be processed; searching for neighboring nodes adjacent to the target trajectory node from the true trajectory, and determining the true coordinates of the target trajectory node based on the coordinates and timestamps of the neighboring nodes and the timestamp of the target trajectory node based on the interpolation algorithm.

[0087] Specifically, during the vehicle positioning process, the true value trajectory can be obtained at the same time, for example, the RTK trajectory can be obtained at the same time. During the relative positioning accuracy evaluation process, for each trajectory node of the positioning trajectory to be processed, the neighboring nodes adjacent to the trajectory node can be selected from each trajectory node of the RTK trajectory for interpolation operation to determine the true value coordinates of the trajectory node. Figure 3 As shown, for the trajectory node 301 in the positioning trajectory to be processed, the corresponding neighboring nodes (node ​​302 and node 303) can be determined from the trajectory nodes of the RTK trajectory, and then the true coordinates of node 301 can be determined based on the interpolation algorithm according to the timestamp of node 301, the timestamp and coordinates of node 302, and the timestamp and coordinates of node 303. After determining the true coordinates of the target trajectory node, the coordinates of the feature point in the world coordinate system of the vector semantic map and the true coordinates of the target trajectory node can be converted to the same target coordinate system, such as the inertial navigation coordinate system, and then the difference between the coordinates of the feature point and the true coordinates in the target coordinate system is obtained as the target relative distance, and then the difference between the target true distance and the target relative distance can be determined as the relative positioning error, that is, the relative distance error of the semantic elements in the vector semantic map.

[0088] 205. Determine the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the multiple semantic elements.

[0089] In this embodiment, the mean square error or average value of the corresponding error results can be calculated for the same type of semantic elements to determine the relative positioning trajectory. Optionally, the corresponding percentage of positioning errors can be selected according to different confidence requirements, and the mean and covariance of these can be calculated. Exemplarily, the error results corresponding to each semantic element can be sorted. Taking the error values ​​from small to large as an example, the first 60% of the error values ​​can be selected, and the covariance or mean of the confidence interval can be calculated as the relative positioning accuracy corresponding to the confidence interval.

[0090] In some embodiments, in order to more accurately test the relative positioning accuracy, the relative positioning accuracy can be determined in segments for different scenes (e.g., highway scenes, canyon scenes, tunnel scenes, urban scenes, mountain scenes, etc.). Determining the relative positioning accuracy of the positioning trajectory to be processed based on the error results corresponding to the multiple semantic elements may include: dividing the positioning trajectory to be processed into multiple segments based on different scenes to obtain multiple sub-trajectories; for each sub-trajectory, determining the error result corresponding to the sub-trajectory, and determining the relative positioning accuracy of the sub-trajectory based on the error result.

[0091] The positioning accuracy test method provided in this embodiment uses the laser point cloud as the true value to determine the error result of the distance between the vehicle and the semantic elements around the driving trajectory, thereby obtaining the relative positioning accuracy of the vehicle positioning trajectory.

[0092] There are many ways to calculate the target true value distance. For example, different true value distance calculation methods can be used for different types of semantic elements.

[0093] In some embodiments, for an aerial semantic element, the determining of the target true value distance between the target trajectory node corresponding to the feature point and the first homonymous point corresponding to the feature point may include: if the semantic element is an aerial semantic element, determining the target true value distance according to the coordinates of the first homonymous point corresponding to the feature point. Optionally, the determining of the target true value distance between the target vehicle and the aerial semantic element according to the coordinates of the first homonymous point corresponding to the feature point of the aerial semantic element may include: if the aerial semantic element is a planar semantic element, and the plane where the aerial semantic element is located is perpendicular to the driving direction of the target vehicle, determining the longitudinal coordinate component in the coordinates of the first homonymous point corresponding to the feature point of the aerial semantic element as the longitudinal true value distance between the target trajectory node corresponding to the feature point and the first homonymous point corresponding to the feature point; the determining of the error result corresponding to the semantic element based on the target true value distance and the target measurement distance may include: determining the longitudinal error corresponding to the semantic element based on the longitudinal true value distance and the target measurement distance.

[0094] Specifically, for aerial semantic elements, such as air traffic signs. The coordinates of the first homonymous point in the radar coordinate system can be converted to the target coordinate system, such as the inertial navigation coordinate system, and then based on the shape (e.g., planar, strip, point) and setting method (e.g., parallel to the vehicle's travel direction, vertical setting) of the aerial semantic elements, the target true value distance is determined based on the different components of the coordinates of the first homonymous point in the target coordinate system. Optionally, for aerial semantic elements that are planar and whose planes are perpendicular to the vehicle's travel direction, the longitudinal true value distance can be determined based on the longitudinal component of the coordinates of the first homonymous point, and then the longitudinal error is determined based on the longitudinal true value distance and the longitudinal component of the target measurement distance. For aerial semantic elements that are planar and whose planes are parallel to the vehicle's travel direction, the lateral error can be determined based on the lateral component of the coordinates of the first homonymous point. The above algorithm has high computational efficiency and saves computing power.

[0095] In some embodiments, for ground semantic elements, determining the target true value distance between the target trajectory node corresponding to the feature point and the first homonymous point corresponding to the feature point may include: if the semantic element is a ground semantic element, determining the ground plane equation corresponding to the ground semantic element; according to the ground plane equation, converting the second homonymous point corresponding to the feature point of the ground semantic element from a two-dimensional feature point to a three-dimensional feature point, and determining the target true value distance between the target trajectory node corresponding to the feature point and the first homonymous point corresponding to the feature point according to the coordinates of the three-dimensional feature point. Optionally, determining the target true value distance between the target vehicle and the ground semantic element according to the coordinates of the three-dimensional feature point may include: if the ground plane semantic element is a linear semantic element and the ground semantic element is parallel to the driving direction of the target vehicle, determining the lateral true value distance between the target trajectory node corresponding to the feature point and the first homonymous point corresponding to the feature point according to the lateral coordinates of the coordinates of the three-dimensional feature point; determining the error result corresponding to the semantic element based on the target true value distance and the target measurement distance may include: determining the lateral error corresponding to the semantic element based on the lateral true value distance and the target measurement distance.

[0096] Specifically, the semantic positioning process is a process of restoring the 2D semantic elements of the phase plane to three-dimensional space, and performing 3D registration of the restored 3D semantic elements with the map. Therefore, the ranging accuracy of 2D elements to 3D will directly affect the registration accuracy, and then determine the positioning accuracy. For ground semantic elements, a more accurate ground plane equation can be first obtained based on the laser point cloud group around the semantic elements. There are many ways to determine the ground plane equation. The form of the equation group of the plane can be determined by three points, or by fitting, etc. This embodiment does not limit this. After determining the ground plane equation, the second same-name point can be converted from a 2D point to a 3D point based on the ground plane equation, and the target true value distance between the target trajectory point and the corresponding first same-name point can be determined based on the coordinates of the 3D point. For the shape characteristics of the ground semantic elements, the corresponding error can be determined based on the different components of the 3D point coordinates. For example, strip-shaped semantic elements parallel to the vehicle's driving direction (e.g., lane lines) can record lateral components to determine lateral errors, and semantic elements with vertices (e.g., dotted lines and arrows) can obtain lateral components (for determining lateral errors) and longitudinal coordinate values ​​(for determining longitudinal errors) based on the 3D point coordinates of the vertices.

[0097] Figure 4 This is a schematic diagram of the structure of the positioning accuracy test device provided in the embodiment of the present application. Figure 4 As shown, the positioning accuracy testing device 40 includes: an acquisition module 401 and a processing module 402 .

[0098] The acquisition module 401 is used to acquire a target point cloud frame corresponding to a plurality of feature points, and first points of the same name corresponding to the plurality of feature points in the target point cloud frame; the plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to the positioning trajectory to be processed;

[0099] The acquisition module 401 is further used to acquire a target image frame corresponding to the plurality of feature points, and second points of the same name corresponding to the plurality of feature points in the target image frame, wherein the target image frame is collected during the driving process of the target vehicle based on the positioning trajectory to be processed;

[0100] The acquisition module 401 is further used to acquire target trajectory nodes corresponding to the plurality of feature points in the positioning trajectory to be processed;

[0101] The processing module 402 is used to determine, for each feature point of a semantic element, a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determine a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determine an error result corresponding to the semantic element based on the target true value distance and the target measured distance;

[0102] The processing module 402 is further configured to determine the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the plurality of semantic elements.

[0103] The positioning accuracy testing equipment provided in the embodiment of the present application determines the error result of the distance between the vehicle and the semantic elements around the driving trajectory by taking the laser point cloud as the true value, and then obtains the relative positioning accuracy of the vehicle positioning trajectory.

[0104] The positioning accuracy testing device provided in the embodiment of the present application can be used to execute the above-mentioned method embodiment. Its implementation principle and technical effects are similar, and this embodiment will not be repeated here.

[0105] Figure 5 This is a structural block diagram of a positioning accuracy test device provided in an embodiment of the present application. The device may be a computer, tablet, mobile phone, server, or other device with computing capabilities.

[0106] The device 50 may include one or more of the following components: a processing component 501 , a memory 502 , a power component 503 , a multimedia component 504 , an audio component 505 , an input / output (I / O) interface 506 , a sensor component 507 , and a communication component 508 .

[0107] The processing component 501 generally controls the overall operation of the device 50, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 501 may include one or more processors 509 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 501 may include one or more modules to facilitate the interaction between the processing component 501 and other components. For example, the processing component 501 may include a multimedia module to facilitate the interaction between the multimedia component 504 and the processing component 501.

[0108] The memory 502 is configured to store various types of data to support operations on the device 50. Examples of such data include instructions for any application or method operating on the device 50, contact data, phone book data, messages, pictures, videos, etc. The memory 502 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0109] The power supply component 503 provides power to the various components of the device 50. The power supply component 503 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device 50.

[0110] The multimedia component 504 includes a screen that provides an output interface between the device 50 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 504 includes a front camera and / or a rear camera. When the device 50 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0111] The audio component 505 is configured to output and / or input audio signals. For example, the audio component 505 includes a microphone (MIC), and when the device 50 is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 502 or sent via the communication component 508. In some embodiments, the audio component 505 also includes a speaker for outputting audio signals.

[0112] I / O interface 506 provides an interface between processing component 501 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0113] The sensor assembly 507 includes one or more sensors for providing various aspects of status assessment for the device 50. For example, the sensor assembly 507 can detect the open / closed state of the device 50, the relative positioning of components, such as the display and keypad of the device 50, and the sensor assembly 507 can also detect the position change of the device 50 or a component of the device 50, the presence or absence of user contact with the device 50, the orientation or acceleration / deceleration of the device 50, and the temperature change of the device 50. The sensor assembly 507 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 507 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 507 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0114] The communication component 508 is configured to facilitate wired or wireless communication between the device 50 and other devices. The device 50 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 508 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 508 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0115] In an exemplary embodiment, the device 50 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above-mentioned methods.

[0116] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 502 including instructions, and the instructions can be executed by a processor 509 of the device 50 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0117] The computer-readable storage medium mentioned above can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0118] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0119] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0120] An embodiment of the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the positioning accuracy testing method performed by the positioning accuracy testing device as described above.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A positioning accuracy testing method, characterized in that: The method includes: Acquire a target point cloud frame corresponding to a plurality of feature points, and first points of the same name corresponding to the plurality of feature points in the target point cloud frame; the plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to the positioning trajectory to be processed; Acquire a target image frame corresponding to a plurality of the feature points, and second points of the same name corresponding to the plurality of the feature points in the target image frame, wherein the target image frame is collected during the driving process of the target vehicle based on the positioning trajectory to be processed; Acquire target trajectory nodes respectively corresponding to a plurality of the feature points in the positioning trajectory to be processed; For each feature point of a semantic element, determining a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determining a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determining an error result corresponding to the semantic element based on the target true value distance and the target measured distance; Determining the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the plurality of semantic elements; The step of obtaining a target point cloud frame corresponding to a plurality of feature points includes: Obtaining a positioning trajectory to be processed, a vector semantic map corresponding to the positioning trajectory to be processed, and a corresponding laser point cloud frame; Selecting feature points corresponding to a plurality of semantic elements respectively from the vector semantic map; the plurality of semantic elements are located within a preset range of the positioning trajectory to be processed; A plurality of target point cloud frames corresponding to the feature points are selected from the laser point cloud frames.

2. The method according to claim 1, characterized in that: The step of selecting a plurality of target point cloud frames corresponding to the feature points from the laser point cloud frames comprises: Acquire sensor data corresponding to the positioning trajectory to be processed; Based on the positioning trajectory to be processed and the sensor data, the coordinates of each point cloud in the laser point cloud frame are converted to the same coordinate system to obtain a corrected laser point cloud frame; A plurality of target point cloud frames corresponding to the feature points are selected from the corrected laser point cloud frames.

3. The method according to claim 1, characterized in that The determining of the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point includes: Obtaining a distance measurement result corresponding to a second point of the same name corresponding to the feature point; the distance measurement result is a distance between the second point of the same name generated in the process of generating the positioning trajectory to be processed and the target trajectory node; The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes: The difference between the distance measurement result and the target true value distance is determined as the distance measurement error corresponding to the semantic element.

4. The method according to claim 1, characterized in that The determining of the target measurement distance between the target trajectory node corresponding to the feature point and the second point of the same name corresponding to the feature point includes: Determine the true value coordinates of the target trajectory node corresponding to the feature point; The coordinates of the feature points and the true value coordinates of the corresponding target trajectory nodes are converted into a target coordinate system, and the target relative distance between the feature points of the semantic elements and the target trajectory nodes is calculated; The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes: The difference between the target true value distance and the target relative distance is determined as the relative positioning error corresponding to the semantic element.

5. The method according to claim 4, characterized in that Determining the true value coordinates of the target trajectory node includes: Obtaining a true value trajectory corresponding to the positioning trajectory to be processed; A neighboring node adjacent to the target trajectory node is searched from the true trajectory, and based on an interpolation algorithm, the true coordinates of the target trajectory node are determined according to the coordinates and timestamps of the neighboring nodes and the timestamp of the target trajectory node.

6. The method according to any one of claims 1 to 5, characterized in that: The determining of the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point includes: If the semantic element is an aerial semantic element, the target true value distance is determined according to the coordinates of the first point of the same name corresponding to the feature point.

7. The method according to claim 6, characterized in that The determining the target true value distance between the target vehicle and the aerial semantic element according to the coordinates of the first same-name point corresponding to the feature point of the aerial semantic element comprises: If the aerial semantic element is a planar semantic element, and the plane where the aerial semantic element is located is perpendicular to the driving direction of the target vehicle, the longitudinal coordinate component in the coordinates of the first same-name point corresponding to the feature point of the aerial semantic element is determined as the longitudinal true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point; The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes: Based on the longitudinal true value distance and the target measurement distance, a longitudinal error corresponding to the semantic element is determined.

8. The method according to any one of claims 1 to 5, characterized in that: The determining of the target true value distance between the target trajectory node corresponding to the feature point and the first point of the same name corresponding to the feature point includes: If the semantic element is a ground semantic element, determining a ground plane equation corresponding to the ground semantic element according to a second point of the same name corresponding to the feature point; According to the ground plane equation, the second same-name point corresponding to the feature point of the ground semantic element is converted from a two-dimensional feature point to a three-dimensional feature point, and according to the coordinates of the three-dimensional feature point, the target true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point is determined.

9. The method according to claim 8, characterized in that Determining the target true value distance between the target vehicle and the ground semantic element according to the coordinates of the three-dimensional feature point includes: If the ground plane semantic element is a linear semantic element, and the ground plane semantic element is parallel to the driving direction of the target vehicle, then determining the lateral true value distance between the target trajectory node corresponding to the feature point and the first same-name point corresponding to the feature point according to the lateral coordinate of the coordinate of the three-dimensional feature point; The determining, based on the target true value distance and the target measured distance, an error result corresponding to the semantic element includes: Based on the lateral true value distance and the target measurement distance, a lateral error corresponding to the semantic element is determined.

10. The method according to any one of claims 1 to 5, characterized in that: The step of determining the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the plurality of semantic elements comprises: Based on different scenarios, the positioning trajectory to be processed is divided into multiple segments to obtain multiple sub-trajectories; For each sub-trajectory, an error result corresponding to the sub-trajectory is determined, and a relative positioning accuracy of the sub-trajectory is determined according to the error result.

11. A positioning accuracy testing device, characterized in that: include: An acquisition module, used to acquire a target point cloud frame corresponding to a plurality of feature points, and first points of the same name corresponding to the plurality of feature points in the target point cloud frame; The plurality of feature points are feature points corresponding to a plurality of semantic elements in a vector semantic map corresponding to the positioning trajectory to be processed; The acquisition module is further used to acquire target image frames corresponding to the plurality of feature points, and second points of the same name corresponding to the plurality of feature points in the target image frames, wherein the target image frames are collected during the driving process of the target vehicle based on the positioning trajectory to be processed; The acquisition module is further used to acquire target trajectory nodes respectively corresponding to the plurality of feature points in the positioning trajectory to be processed; a processing module, configured to determine, for each feature point of a semantic element, a target true value distance between a target trajectory node corresponding to the feature point and a first same-name point corresponding to the feature point, determine a target measured distance between a target trajectory node corresponding to the feature point and a second same-name point corresponding to the feature point, and determine an error result corresponding to the semantic element based on the target true value distance and the target measured distance; The processing module is further used to determine the relative positioning accuracy of the positioning trajectory to be processed according to the error results corresponding to the plurality of semantic elements; The acquisition module is specifically used to obtain a positioning trajectory to be processed, a vector semantic map corresponding to the positioning trajectory to be processed, and a corresponding laser point cloud frame; select feature points corresponding to multiple semantic elements from the vector semantic map; multiple semantic elements are located within a preset range of the positioning trajectory to be processed; and select target point cloud frames corresponding to multiple feature points from the laser point cloud frame.

12. An electronic device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the positioning accuracy testing method as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the processor executes the computer-executable instructions, the positioning accuracy testing method according to any one of claims 1 to 10 is implemented.

Images