Positioning reliability testing methods and related equipment

Abstract

This application relates to the field of intelligent driving technology, specifically providing a positioning reliability detection method and related equipment. The method includes: acquiring a first positioning result of a target vehicle; determining a first element image based on the first positioning result and a high-precision map; and performing reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result. Using this application embodiment, the reliability of the positioning result of autonomous vehicles can be detected, which helps to reduce safety hazards in autonomous driving.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and in particular to a positioning reliability detection method and related equipment. Background Technology

[0002] Autonomous vehicles heavily rely on positioning technology. Positioning is a prerequisite for other functions in autonomous driving; lost or incorrect positioning will render autonomous vehicles inoperable and could even lead to major accidents. However, most current high-precision positioning methods employ fusion positioning, which, by adding algorithms (such as visual odometry and laser odometry), cannot provide a reliable assessment of positioning performance. Currently, sensor-based positioning equipment can generally determine the sensor's positioning performance based on observations, providing the positioning error covariance. This covariance can reflect the reliability of the positioning result to some extent, but its confidence level is low. Since the control and perception algorithms in autonomous driving technology depend on the positioning results, inaccurate positioning will severely impact these algorithms, creating safety hazards for autonomous driving. Summary of the Invention

[0003] This application discloses a positioning reliability detection method and related equipment, which can perform reliability detection on the positioning results of autonomous vehicles, and helps to reduce safety hazards in autonomous driving.

[0004] The first aspect of this application discloses a positioning reliability detection method applied to an in-vehicle terminal. The method includes: acquiring a first positioning result of a target vehicle; determining a first element image based on the first positioning result and a high-precision map; and performing reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result.

[0005] As can be seen, in this embodiment, after obtaining the first positioning result of the target vehicle, the vehicle terminal determines the first element image based on the first positioning result and the high-precision map, and then performs a reliability test on the first positioning result based on the first element image to obtain the first detection result of the first positioning result. Because the first positioning result is converted into a binary image based on the high-precision map, the first positioning result is compared with the high-precision map to detect whether the first positioning result is reliable. If the first positioning result is reliable, it can be used, effectively reducing the harm caused by unreliable positioning and helping to reduce safety hazards in autonomous driving.

[0006] In one exemplary embodiment, the above-mentioned reliability detection of the first localization result based on the first feature image to obtain the first detection result of the first localization result includes: inputting the first feature image into a target detection model for processing to obtain the first detection result of the first localization result; wherein the target detection model is implemented based on the Visual Geometry Group (VGG) architecture.

[0007] As can be seen, in this embodiment, the first feature image obtained by converting the first positioning result is input into the target detection model implemented based on the VGG architecture. The target detection model is used to judge the first feature image and detect whether the first feature image is a feature image obtained by converting a normal positioning result or a feature image obtained by converting an abnormal positioning result, thereby detecting whether the first positioning result is reliable.

[0008] In an exemplary embodiment, before inputting the first feature image into the target detection model for processing to obtain the first detection result of the first localization result, the method further includes: acquiring N second localization results and M third localization results of the vehicle, wherein the second localization results are historical localization results with normal localization, the third localization results are historical localization results with abnormal localization, and N and M are positive integers; determining N second feature images based on the N second localization results and the high-precision map, and determining M third feature images based on the M third localization results and the high-precision map, wherein the N second localization results correspond one-to-one with the N second feature images, and the M third localization results correspond one-to-one with the M third feature images; training the initial detection model using the N second feature images and the M third feature images to obtain the target detection model; wherein the initial detection model is implemented based on the VGG architecture.

[0009] As can be seen, in this embodiment, before using the target detection model to discriminate the first element image, a large number of historical vehicle positioning results are obtained; based on the high-precision map, the second positioning results with normal positioning are converted into second element images as normal positioning category images, and the third positioning results with abnormal positioning are converted into third element images as abnormal positioning category images; these second and third element images are used to train the initial detection model based on the VGG architecture to obtain the target detection model, which is beneficial for using the target detection model to perform reliability detection on the real-time first positioning results of the target vehicle.

[0010] In one exemplary embodiment, a target element image is determined based on the target positioning result and a high-precision map through the following steps: obtaining map element point data within a preset area from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map element point data includes multiple map element points; converting the first coordinates of each of the multiple map element points into second coordinates, wherein the first coordinates of each map element point are the coordinates of the map element point in the world coordinate system, and the second coordinates are the coordinates of the map element point in the second coordinate system, wherein the second coordinate system is a coordinate system with the positioning coordinates in the target positioning result as the origin; converting the second coordinates of each map element point into image coordinates according to the camera intrinsic parameters of the vehicle terminal to obtain the target element image, wherein the pixel coordinates of the target element image are the image coordinates; wherein, if the target positioning result is the first positioning result, then the target element image is the first element image; if the target positioning result is the second positioning result, then the target element image is the second element image; if the target positioning result is the third positioning result, then the target element image is the third element image.

[0011] As can be seen, in this embodiment, map feature point data within a preset area is first obtained from a high-precision map. The preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points. Then, the first coordinates of each of these multiple map feature points are converted into second coordinates. The first coordinates of each map feature point are its coordinates in the world coordinate system, and the second coordinates are its coordinates in the second coordinate system, which is a coordinate system with the positioning coordinates in the target positioning result as its origin. Next, based on the camera intrinsic parameters of the vehicle terminal, the second coordinates of each map feature point are converted into image coordinates to obtain the target feature image. The pixel coordinates of the target feature image are the image coordinates, thereby converting the positioning result into a feature image. This allows for comparison between the positioning result and the high-precision map, which is beneficial for detecting the reliability of the positioning result. If the target positioning result is the first positioning result, then the target feature image is the first feature image, thus enabling the detection of the reliability of the real-time positioning result of the target vehicle. If the target positioning result is the second positioning result, then the target element image is the second element image; if the target positioning result is the third positioning result, then the target element image is the third element image. In this way, the element images corresponding to the vehicle's historical positioning results can be obtained. The initial detection model can be trained using these element images corresponding to the historical positioning results to obtain a target detection model for detecting the reliability of real-time positioning results.

[0012] In one exemplary embodiment, the first coordinates are converted into the second coordinates according to the following formula:

[0013]

[0014] Among them, (X) w ,Y w Z w (X) is the first coordinate mentioned above. c ,Y c Z c R is the second coordinate mentioned above, R is the rotation matrix of the three degrees of freedom of the attitude in the target positioning result mentioned above, and T is the translation matrix of the positioning coordinate in the target positioning result mentioned above.

[0015] As can be seen, in this embodiment, the first coordinates of the map feature points are transformed into the second coordinates by the rotation matrix of the three degrees of freedom of the attitude in the target positioning result and the translation matrix of the positioning coordinates in the target positioning result. Since the second coordinates are the coordinates of the map feature points in the coordinate system with the positioning coordinates in the target positioning result as the origin, the coordinates of the map feature points in the high-precision map are transformed into the coordinate system with the vehicle positioning coordinates as the origin, which is beneficial for detecting whether the vehicle positioning result is reliable.

[0016] In one exemplary implementation, the second coordinates described above are converted into image coordinates according to the following formula:

[0017]

[0018] Where (u,v) are the image coordinates mentioned above, f x f y Let u0 and v0 be the focal length of the camera in the aforementioned vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

[0019] As can be seen, in this embodiment, the second coordinates are converted into image coordinates by the focal length of the camera of the vehicle terminal, and the target element image corresponding to the target positioning result is obtained. The pixel coordinates of the target element image are the image coordinates obtained by converting the second coordinates, so as to compare the target positioning result with the high-precision map, which is beneficial to detect the reliability of the target positioning result.

[0020] In one exemplary embodiment, the method further includes: if the first positioning result comes directly from the positioning sensor, then obtaining the positioning error covariance determined by the positioning sensor for the first positioning result; and determining a second detection result of the first positioning result based on the first detection result and the positioning error covariance.

[0021] As can be seen, in this embodiment, when the positioning sensor directly provides the first positioning result, the positioning sensor will provide the positioning error covariance corresponding to the first positioning result. This positioning error covariance can reflect the reliability of the first positioning result. Based on this, the reliability of the first positioning result can be further evaluated by combining the first detection result, which can further improve the confidence level of the reliability of the first positioning result.

[0022] In one exemplary embodiment, the method further includes: if the first detection result and the second detection result indicate that the first positioning result is normal, then the first positioning result is used for autonomous driving; if the first detection result and / or the second detection result indicate that the first positioning result is abnormal, then an alert is issued to disable autonomous driving.

[0023] Therefore, in this embodiment, when the first positioning result is detected as normal, it indicates that the first positioning result is reliable and can be used for autonomous driving of the autonomous vehicle; when the first positioning result is detected as abnormal, it indicates that the first positioning result is unreliable, and it can warn to shut down the autonomous driving of the autonomous vehicle and allow manual takeover, thereby reducing the safety hazards caused by unreliable positioning.

[0024] A second aspect of this application discloses a positioning reliability detection device applied to an in-vehicle terminal. The device includes: an acquisition unit for acquiring a first positioning result of a target vehicle; a determination unit for determining a first element image based on the first positioning result and a high-precision map; and a detection unit for performing reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result.

[0025] In one exemplary embodiment, the detection unit is configured to: input the first element image into a target detection model for processing to obtain a first detection result of the first localization result; wherein the target detection model is implemented based on the VGG architecture.

[0026] In one exemplary embodiment, the apparatus further includes a training unit, configured to: acquire N second positioning results and M third positioning results of the vehicle, wherein the second positioning results are historical positioning results with normal positioning, the third positioning results are historical positioning results with abnormal positioning, and N and M are positive integers; determine N second feature images based on the N second positioning results and a high-precision map, and determine M third feature images based on the M third positioning results and the high-precision map, wherein the N second positioning results correspond one-to-one with the N second feature images, and the M third positioning results correspond one-to-one with the M third feature images; train an initial detection model using the N second feature images and the M third feature images to obtain the target detection model; wherein the initial detection model is implemented based on the VGG architecture.

[0027] In an exemplary embodiment, the determining unit is configured to: acquire map feature point data within a preset area from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points; convert the first coordinates of each of the multiple map feature points into second coordinates, wherein the first coordinates of each map feature point are the coordinates of the map feature point in the world coordinate system, and the second coordinates are the coordinates of the map feature point in the second coordinate system, the second coordinate system being a coordinate system with the positioning coordinates in the target positioning result as the origin; convert the second coordinates of each map feature point into image coordinates according to the camera intrinsic parameters of the vehicle terminal to obtain the target feature image, wherein the pixel coordinates of the target feature image are the image coordinates; wherein if the target positioning result is the first positioning result, then the target feature image is the first feature image; if the target positioning result is the second positioning result, then the target feature image is the second feature image; if the target positioning result is the third positioning result, then the target feature image is the third feature image.

[0028] In one exemplary embodiment, the first coordinates are converted into the second coordinates according to the following formula:

[0029]

[0030] Among them, (X) w ,Y w Z w (X) is the first coordinate mentioned above. c ,Y c Z cR is the second coordinate mentioned above, R is the rotation matrix of the three degrees of freedom of the attitude in the target positioning result mentioned above, and T is the translation matrix of the positioning coordinate in the target positioning result mentioned above.

[0031] In one exemplary implementation, the second coordinates described above are converted into image coordinates according to the following formula:

[0032]

[0033] Where (u,v) are the image coordinates mentioned above, f x f y Let u0 and v0 be the focal length of the camera in the aforementioned vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

[0034] In one exemplary embodiment, the detection unit is further configured to: if the first positioning result comes directly from the positioning sensor, obtain the positioning error covariance determined by the positioning sensor for the first positioning result; and determine a second detection result of the first positioning result based on the first detection result and the positioning error covariance.

[0035] In one exemplary embodiment, the above-mentioned device further includes a processing unit, configured to: if the first detection result and the second detection result indicate that the first positioning result is normal, then use the first positioning result for automatic driving; if the first detection result and / or the second detection result indicate that the first positioning result is abnormal, then issue a warning to disable automatic driving.

[0036] A third aspect of this application discloses an in-vehicle terminal, including a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for performing the steps of the method as described in any one of the first aspects above.

[0037] The fourth aspect of this application discloses a chip, characterized in that it includes: a processor for calling and running a computer program from a memory, causing a device equipped with the chip to perform the method described in any of the first aspects above.

[0038] The fifth aspect of this application discloses a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform the method as described in any one of the first aspects above.

[0039] A sixth aspect of this application discloses a computer program product that causes a computer to perform the method as described in any one of the first aspects above.

[0040] A seventh aspect of this application discloses an intelligent vehicle, which includes a control system that performs the method described in any one of the first aspects above to control the driving of the intelligent vehicle. Attached Figure Description

[0041] The accompanying drawings used in the embodiments of this application are described below.

[0042] Figure 1 This is a schematic diagram of the structure of a positioning reliability detection system provided in an embodiment of this application;

[0043] Figure 2 This is a schematic flowchart of a positioning reliability detection method provided in an embodiment of this application;

[0044] Figure 3 This is a schematic diagram of an element image provided in an embodiment of this application;

[0045] Figure 4 This is a schematic diagram of another element image provided in an embodiment of this application;

[0046] Figure 5 This is a schematic diagram illustrating an embodiment of obtaining map feature point data provided in this application.

[0047] Figure 6 This is a schematic diagram of the structure of a positioning reliability detection device provided in an embodiment of this application;

[0048] Figure 7 This is a schematic diagram of the structure of a vehicle-mounted terminal provided in an embodiment of this application. Detailed Implementation

[0049] The embodiments of this application are described below with reference to the accompanying drawings.

[0050] The technical solutions provided in this application can be applied to, but are not limited to, the following scenarios:

[0051] Scenario 1: Positioning is achieved using algorithms, such as multi-sensor fusion. In this scenario, it is difficult to provide a reliability index for the positioning results. However, the technical solution provided in the embodiments of this application can provide a reliable positioning result.

[0052] Scenario 2: Positioning results are directly provided by positioning sensors, such as the Global Positioning System (GPS) and the Inertial Measurement Unit (IMU). In this scenario, GPS and IMU will provide the positioning error covariance corresponding to the positioning results. The positioning error covariance can reflect the reliability index of the positioning results. Combined with the technical solution provided in the embodiments of this application, the confidence level of the positioning results can be further improved.

[0053] To address the challenges of providing reliable positioning results in real time for high-precision fusion positioning achieved through algorithms, and the low confidence level of the positioning error covariance provided by positioning sensors, this application provides a positioning reliability detection system. This system can perform real-time reliability assessment of the positioning results based on the current fusion positioning algorithm, or combine the positioning error covariance provided by the positioning sensors to perform real-time reliability assessment of the positioning results. This provides a criterion for subsequent perception and control algorithms to determine whether to use the positioning results for autonomous driving, effectively addressing potential safety hazards.

[0054] Please see Figure 1 , Figure 1 This is a schematic diagram of a positioning reliability detection system provided in an embodiment of this application. The system includes a projection module, a reliability detection module, and a security module. The projection module combines a high-precision map with the element images obtained from the positioning results. Specifically, the projection module obtains continuous positioning results from an algorithm or directly from a positioning sensor. Based on these positioning results and the high-precision map, it projects map elements onto an image coordinate system to obtain the projected element images. Optionally, the reliability detection module, based on the VGG model, can input the element images projected by the projection module to the reliability detection module for reliability judgment of the positioning results and output the reliability of the positioning results. The security module performs subsequent processing based on the real-time judgment results output by the reliability detection module and provides a warning indicating whether manual intervention is required.

[0055] The aforementioned high-precision maps refer to electronic maps with higher accuracy and more data dimensions. Higher accuracy is reflected in centimeter-level precision, while more data dimensions are reflected in the inclusion of surrounding static information related to traffic, in addition to road information. High-precision maps store a large amount of driving assistance information as structured data, which can be divided into two categories. The first category is road data, such as lane information including lane line location, type, width, slope, and curvature. The second category is information on fixed objects around the lanes, such as traffic signs, traffic lights, lane height restrictions, sewer openings, obstacles, and other road details, including information on elevated objects, guardrails, their number, road edge types, roadside landmarks, and other infrastructure.

[0056] The aforementioned element image refers to the acquisition of map element point data within a preset range of the positioning coordinates on a high-precision map based on the positioning coordinates provided by the algorithm or directly provided by the positioning sensor; then, the coordinates of the map element points in the world coordinate system are converted into coordinates in the local coordinate system with the positioning coordinates as the origin; then, the coordinates of the map element points in the local coordinate system are converted into image coordinates, and the image projected by the map element points according to the image coordinates is the element image.

[0057] exist Figure 1 The described positioning reliability detection system effectively compares the positioning results with a high-precision map, thereby outputting the reliability of the positioning results. This provides a basis for subsequent algorithms to decide whether to continue using the positioning results, which helps improve the safety of autonomous vehicles.

[0058] Please see Figure 2 , Figure 2 This is a flowchart illustrating a positioning reliability detection method provided in an embodiment of this application. The method is applied to an in-vehicle terminal and includes, but is not limited to, the following steps:

[0059] Step 201: Obtain the first location result of the target vehicle.

[0060] The first positioning result mentioned above includes the positioning coordinates of the target vehicle and the attitude information of the target vehicle.

[0061] Specifically, the aforementioned vehicle-mounted terminal can be installed on the target vehicle. The vehicle-mounted terminal can obtain the first positioning result of the target vehicle given by the fusion positioning algorithm in real time, or obtain the first positioning result of the target vehicle directly given by positioning devices such as GPS and IMU sensors in real time.

[0062] Step 202: Determine the first element image based on the first positioning result and the high-precision map.

[0063] The vehicle-mounted terminal can store a high-precision map. After obtaining the first positioning result of the target vehicle, it can obtain the first element image based on the first positioning result and the projection of the high-precision map.

[0064] Specifically, the first element image is obtained by acquiring map element point data within a preset range of the target vehicle's positioning coordinates on a high-precision map based on the target vehicle's positioning coordinates in the first positioning result, and then performing coordinate transformation on the map element point coordinates in the map element point data according to the first positioning result. Therefore, the first positioning result is projected into the high-precision map to obtain the first element image; in other words, the target vehicle is projected into the high-precision map to obtain the first element image.

[0065] For example, map feature point data within a semi-circular area in the forward-looking direction of the target vehicle's positioning coordinates can be obtained, and then projected to obtain the first feature image. At this time, the first feature image is an image in the forward-looking direction of the target vehicle, which is equivalent to projecting the target vehicle into a high-precision map. The image seen from the forward-looking direction of the target vehicle is the first feature image.

[0066] Step 203: Perform a reliability test on the first positioning result based on the first element image to obtain a first detection result for the first positioning result.

[0067] Specifically, the first element image can be input into an image recognition model for recognition, thereby judging whether the first localization result is reliable.

[0068] For example, when the first element image is the view seen from the target vehicle's forward direction, if the position and / or orientation of the map element points in the first element image matches the position and / or orientation of the map element points seen when the vehicle is driving normally, then the first positioning result is reliable; otherwise, the first positioning result is unreliable. For instance, when a vehicle is driving normally on the road, its forward direction and the road lines are parallel. If the road lines in the first element image intersect with the vehicle's forward direction, then the first positioning result is unreliable; or when a vehicle is driving normally on the road, its forward direction points to the end of the road. If the vehicle's forward direction in the first element image points to the sky or another direction, then the first positioning result is unreliable.

[0069] exist Figure 2 In the described method, after acquiring the first positioning result of the target vehicle, the vehicle terminal determines a first element image based on the first positioning result and a high-precision map. Then, it performs a reliability test on the first positioning result based on the first element image to obtain a first detection result for the first positioning result. By converting the first positioning result into a binary image based on the high-precision map, the first positioning result and the high-precision map are compared to detect whether the first positioning result is reliable. If the first positioning result is reliable, it can be used, effectively reducing the harm caused by unreliable positioning and helping to reduce safety hazards in autonomous driving.

[0070] In one exemplary embodiment, the above-mentioned reliability detection of the first localization result based on the first feature image to obtain a first detection result of the first localization result includes: inputting the first feature image into a target detection model for processing to obtain a first detection result of the first localization result; wherein the target detection model is implemented based on a visual geometric group network architecture.

[0071] Specifically, the aforementioned target detection model includes an offline-trained classification model with VGG16 as the backbone network. The first element image can be input into the offline-trained classification model with VGG16 as the backbone network to judge the first element image and determine whether the first element image is an element image output by the normal localization result or an element image output by the abnormal localization result.

[0072] For example, please refer to the following: Figure 3 , Figure 3 This is a schematic diagram of an element image provided in an embodiment of this application. For example... Figure 3 As shown, Figure 3 This is the first element image obtained by converting the target vehicle's initial positioning result when it is normal. If it is input into the aforementioned offline-trained classification model with VGG16 as the backbone network, the offline-trained classification model with VGG16 as the backbone network can distinguish it, for example, distinguishing the position of the map element point in the target vehicle's forward-looking direction from the position of the same map element point in the forward-looking direction when the vehicle is driving normally. When the vehicle is driving normally on the road, the forward-looking direction of the vehicle is parallel to the road lines, and the offline-trained classification model with VGG16 as the backbone network can distinguish it from the position of the target vehicle's forward-looking direction when the vehicle is driving normally. Figure 3 The perspective shown indicates that the target vehicle's forward-looking direction and the road lines are parallel, thus confirming that the first positioning result is reliable.

[0073] For example, please refer to the following: Figure 4 , Figure 4 This is a schematic diagram of another element image provided in an embodiment of this application. For example... Figure 4 As shown, Figure 4 This is the first element image obtained when the target vehicle's initial positioning result is abnormal; if it is input into the aforementioned offline-trained classification model with VGG16 as the backbone network for discrimination, the offline-trained classification model with VGG16 as the backbone network will... Figure 4 The perspective shown indicates that the target vehicle's forward-looking direction intersects with the road line, whereas when the vehicle is driving normally on the road, its forward-looking direction is parallel to the road line, thus determining that the first positioning result is unreliable.

[0074] As can be seen, in this embodiment, the first feature image obtained by converting the first positioning result is input into the target detection model implemented based on the VGG architecture. The target detection model is used to judge the first feature image and detect whether the first feature image is a feature image obtained by converting a normal positioning result or a feature image obtained by converting an abnormal positioning result, thereby detecting whether the first positioning result is reliable.

[0075] In an exemplary embodiment, before inputting the first feature image into the target detection model for processing to obtain the first detection result of the first localization result, the method further includes: acquiring N second localization results and M third localization results of the vehicle, wherein the second localization results are historical localization results with normal localization, the third localization results are historical localization results with abnormal localization, and N and M are positive integers; determining N second feature images based on the N second localization results and the high-precision map, and determining M third feature images based on the M third localization results and the high-precision map, wherein the N second localization results correspond one-to-one with the N second feature images, and the M third localization results correspond one-to-one with the M third feature images; training the initial detection model using the N second feature images and the M third feature images to obtain the target detection model; wherein the initial detection model is implemented based on the visual geometric group network architecture.

[0076] Specifically, the initial detection model described above is a classification model using VGG16 as the backbone network. It can record real vehicle GPS positioning data, project the second feature image generated from the second positioning result with normal positioning as the normal positioning category image, and project the third feature image generated from the third positioning result with abnormal positioning as the abnormal positioning category image. Then, these normal positioning category images and abnormal positioning category images are input into the initial detection model for training, thereby obtaining the target detection model.

[0077] In the training process of the classification model with VGG16 as the backbone network, the ImageNet pre-trained model parameters were used as the initial weights of the network; the training image size was 224*224; the training categories were two, namely normal localization category images and abnormal localization category images; dropout was set to 0.5; and the batch size was 8.

[0078] As can be seen, in this embodiment, before using the target detection model to discriminate the first element image, a large number of historical vehicle positioning results are obtained; based on the high-precision map, the second positioning results with normal positioning are converted into second element images as normal positioning category images, and the third positioning results with abnormal positioning are converted into third element images as abnormal positioning category images; these second and third element images are used to train the initial detection model based on the VGG architecture to obtain the target detection model, which is beneficial for using the target detection model to perform reliability detection on the real-time first positioning results of the target vehicle.

[0079] In one exemplary embodiment, a target element image is determined based on the target positioning result and a high-precision map through the following steps: obtaining map element point data within a preset area from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map element point data includes multiple map element points; converting the first coordinates of each of the multiple map element points into second coordinates, wherein the first coordinates of each map element point are the coordinates of the map element point in the world coordinate system, and the second coordinates are the coordinates of the map element point in the second coordinate system, wherein the second coordinate system is a coordinate system with the positioning coordinates in the target positioning result as the origin; converting the second coordinates of each map element point into image coordinates according to the camera intrinsic parameters of the vehicle terminal to obtain the target element image, wherein the pixel coordinates of the target element image are the image coordinates; wherein, if the target positioning result is the first positioning result, then the target element image is the first element image; if the target positioning result is the second positioning result, then the target element image is the second element image; if the target positioning result is the third positioning result, then the target element image is the third element image.

[0080] Specifically, such as Figure 5 As shown, the target location result P(x,y,z,r,p,y) and attitude (roll,p,yaw) are obtained, and a high-precision map is generated. A semi-circular area with radius γ centered on the target location result P in the vehicle's forward-looking direction is defined as the preset area. Map feature point data map_near within this semi-circular area is obtained, which is the set of map feature points in the world coordinate system for all map features within this semi-circular area. Map features include: utility poles, lane lines, guide lines, etc. γ is a settable value, with a default value of 50m. Then, for the map feature points in map_near, their first coordinate (X,y,z) in the world coordinate system is calculated. w ,Y w Z w Transform to the second coordinate (X) in the local coordinate system with P as the origin. c ,Y c Z c Then, for map feature points in the local coordinate system map_near with P as the origin, the second coordinate (X) of the map feature points is obtained through camera intrinsics. c ,Y c Z c The coordinates are converted into image coordinates (u,v), which is to generate pixel coordinates (u,v), thereby generating the projected image of the target element.

[0081] As can be seen, in this embodiment, map feature point data within a preset area is first obtained from a high-precision map. The preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points. Then, the first coordinates of each of these multiple map feature points are converted into second coordinates. The first coordinates of each map feature point are its coordinates in the world coordinate system, and the second coordinates are its coordinates in the second coordinate system, which is a coordinate system with the positioning coordinates in the target positioning result as its origin. Next, based on the camera intrinsic parameters of the vehicle terminal, the second coordinates of each map feature point are converted into image coordinates to obtain the target feature image. The pixel coordinates of the target feature image are the image coordinates, thereby converting the positioning result into a feature image. This allows for comparison between the positioning result and the high-precision map, which is beneficial for detecting the reliability of the positioning result. If the target positioning result is the first positioning result, then the target feature image is the first feature image, thus enabling the detection of the reliability of the real-time positioning result of the target vehicle. If the target positioning result is the second positioning result, then the target element image is the second element image; if the target positioning result is the third positioning result, then the target element image is the third element image. In this way, the element images corresponding to the vehicle's historical positioning results can be obtained. The initial detection model can be trained using these element images corresponding to the historical positioning results to obtain a target detection model for detecting the reliability of real-time positioning results.

[0082] In one exemplary embodiment, the first coordinates are converted into the second coordinates according to the following formula:

[0083]

[0084] Among them, (X) w ,Y w Z w (X) is the first coordinate mentioned above. c ,Y c Z c R is the second coordinate mentioned above, R is the rotation matrix of the three degrees of freedom of the attitude in the target positioning result mentioned above, and T is the translation matrix of the positioning coordinate in the target positioning result mentioned above.

[0085] As can be seen, in this embodiment, the first coordinates of the map feature points are transformed into the second coordinates by the rotation matrix of the three degrees of freedom of the attitude in the target positioning result and the translation matrix of the positioning coordinates in the target positioning result. Since the second coordinates are the coordinates of the map feature points in the coordinate system with the positioning coordinates in the target positioning result as the origin, the coordinates of the map feature points in the high-precision map are transformed into the coordinate system with the vehicle positioning coordinates as the origin, which is beneficial for detecting whether the vehicle positioning result is reliable.

[0086] In one exemplary implementation, the second coordinates described above are converted into image coordinates according to the following formula:

[0087]

[0088] Where (u,v) are the image coordinates mentioned above, f x f y Let u0 and v0 be the focal length of the camera in the aforementioned vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

[0089] As can be seen, in this embodiment, the second coordinates are converted into image coordinates by the focal length of the camera of the vehicle terminal, and the target element image corresponding to the target positioning result is obtained. The pixel coordinates of the target element image are the image coordinates obtained by converting the second coordinates, so as to compare the target positioning result with the high-precision map, which is beneficial to detect the reliability of the target positioning result.

[0090] In one exemplary embodiment, the method further includes: if the first positioning result comes directly from the positioning sensor, then obtaining the positioning error covariance determined by the positioning sensor for the first positioning result; and determining a second detection result of the first positioning result based on the first detection result and the positioning error covariance.

[0091] Specifically, current positioning devices such as GPS and IMU sensors can generally determine the sensor's positioning performance based on observed values, providing a positioning error covariance. This covariance can reflect the reliability of the positioning result to some extent, but its confidence level is not high. Therefore, when the first positioning result is directly provided by the positioning sensor, based on the positioning error covariance determined by the positioning sensor for the first positioning result, the first detection result of the first positioning result can be combined with the first detection result of the first element image to further detect the first positioning result, thereby obtaining a second detection result of the first positioning result, thus improving the confidence level of the reliability of the first positioning result.

[0092] As can be seen, in this embodiment, when the positioning sensor directly provides the first positioning result, the positioning sensor will provide the positioning error covariance corresponding to the first positioning result. This positioning error covariance can reflect the reliability of the first positioning result. Based on this, the reliability of the first positioning result can be further evaluated by combining the first detection result, which can further improve the confidence level of the reliability of the first positioning result.

[0093] In one exemplary embodiment, the method further includes: if the first detection result and the second detection result indicate that the first positioning result is normal, then the first positioning result is used for autonomous driving; if the first detection result and / or the second detection result indicate that the first positioning result is abnormal, then an alert is issued to disable autonomous driving.

[0094] Specifically, this can be described in conjunction with the aforementioned positioning reliability detection system. When the reliability detection module outputs a normal first positioning result, the safety module determines that the first positioning result is reliable, and the subsequent algorithm module can continue to use the first positioning result for autonomous driving. When the reliability detection module outputs an abnormal first positioning result, the safety module determines that the first positioning result is unreliable, issues an alarm, and reminds the user to turn off autonomous driving and take over manually.

[0095] Therefore, in this embodiment, when the first positioning result is detected as normal, it indicates that the first positioning result is reliable and can be used for autonomous driving of the autonomous vehicle; when the first positioning result is detected as abnormal, it indicates that the first positioning result is unreliable, and it can warn to shut down the autonomous driving of the autonomous vehicle and allow manual takeover, thereby reducing the safety hazards caused by unreliable positioning.

[0096] The methods of the embodiments of this application have been described in detail above, and the apparatus of the embodiments of this application is provided below.

[0097] Please see Figure 6 , Figure 6 This is a schematic diagram of a positioning reliability detection device 600 provided in an embodiment of this application, which is applied to an in-vehicle terminal. The positioning reliability detection device 600 may include an acquisition unit 601, a determination unit 602, and a detection unit 603, wherein the detailed description of each unit is as follows:

[0098] Acquisition unit 601 is used to acquire the first positioning result of the target vehicle;

[0099] The determining unit 602 is used to determine the first element image based on the first positioning result and the high-precision map mentioned above;

[0100] The detection unit 603 is used to perform reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result.

[0101] In one exemplary embodiment, the detection unit 603 is configured to: input the first element image into a target detection model for processing to obtain a first detection result of the first localization result; wherein the target detection model is implemented based on the VGG architecture.

[0102] In one exemplary embodiment, the apparatus further includes a training unit, configured to: acquire N second positioning results and M third positioning results of the vehicle, wherein the second positioning results are historical positioning results with normal positioning, the third positioning results are historical positioning results with abnormal positioning, and N and M are positive integers; determine N second feature images based on the N second positioning results and a high-precision map, and determine M third feature images based on the M third positioning results and the high-precision map, wherein the N second positioning results correspond one-to-one with the N second feature images, and the M third positioning results correspond one-to-one with the M third feature images; train an initial detection model using the N second feature images and the M third feature images to obtain the target detection model; wherein the initial detection model is implemented based on the VGG architecture.

[0103] In one exemplary embodiment, the determining unit 602 is configured to: acquire map feature point data within a preset area from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points; convert the first coordinates of each of the multiple map feature points into second coordinates, wherein the first coordinates of each map feature point are the coordinates of the map feature point in the world coordinate system, and the second coordinates are the coordinates of the map feature point in the second coordinate system, wherein the second coordinate system is a coordinate system with the positioning coordinates in the target positioning result as the origin; convert the second coordinates of each map feature point into image coordinates according to the camera intrinsic parameters of the vehicle terminal to obtain the target feature image, wherein the pixel coordinates of the target feature image are the image coordinates; wherein if the target positioning result is the first positioning result, then the target feature image is the first feature image; if the target positioning result is the second positioning result, then the target feature image is the second feature image; if the target positioning result is the third positioning result, then the target feature image is the third feature image.

[0104] In one exemplary embodiment, the first coordinates are converted into the second coordinates according to the following formula:

[0105]

[0106] Among them, (X) w ,Y w Z w (X) is the first coordinate mentioned above. c ,Y c Z cR is the second coordinate mentioned above, R is the rotation matrix of the three degrees of freedom of the attitude in the target positioning result mentioned above, and T is the translation matrix of the positioning coordinate in the target positioning result mentioned above.

[0107] In one exemplary implementation, the second coordinates described above are converted into image coordinates according to the following formula:

[0108]

[0109] Where (u,v) are the image coordinates mentioned above, f x f y Let u0 and v0 be the focal length of the camera in the aforementioned vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

[0110] In one exemplary embodiment, the detection unit 603 is further configured to: if the first positioning result comes directly from the positioning sensor, obtain the positioning error covariance determined by the positioning sensor for the first positioning result; and determine a second detection result of the first positioning result based on the first detection result and the positioning error covariance.

[0111] In one exemplary embodiment, the above-mentioned device further includes a processing unit, configured to: if the first detection result and the second detection result indicate that the first positioning result is normal, then use the first positioning result for automatic driving; if the first detection result and / or the second detection result indicate that the first positioning result is abnormal, then issue a warning to disable automatic driving.

[0112] It should be noted that the implementation of each unit can also be referenced accordingly. Figure 2 The corresponding description of the method embodiments shown is provided below. Of course, the positioning reliability detection device 600 provided in this application embodiment includes, but is not limited to, the above-described unit modules. For example, the positioning reliability detection device 600 may also include a storage unit 604, which can be used to store the program code and data of the positioning reliability detection device 600.

[0113] exist Figure 6 In the described positioning reliability detection device 600, after acquiring the first positioning result of the target vehicle, a first element image is determined based on the first positioning result and a high-precision map. Then, the reliability of the first positioning result is detected based on the first element image to obtain a first detection result of the first positioning result. Because the first positioning result is converted into a binary image based on the high-precision map, the first positioning result is compared with the high-precision map to detect whether the first positioning result is reliable. If the first positioning result is reliable, it can be used, effectively reducing the harm caused by unreliable positioning and helping to reduce safety hazards in autonomous driving.

[0114] Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of a vehicle terminal 710 provided in an embodiment of this application. The vehicle terminal 710 includes a processor 711, a memory 712 and a communication interface 713. The processor 711, the memory 712 and the communication interface 713 are interconnected through a bus 714.

[0115] The memory 712 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and is used for related computer programs and data. The communication interface 713 is used for receiving and sending data.

[0116] The processor 711 can be one or more central processing units (CPUs). When the processor 711 is a CPU, the CPU can be a single-core CPU or a multi-core CPU.

[0117] The processor 711 in the vehicle terminal 710 is used to read the computer program code stored in the memory 712 and perform the following operations: obtain the first positioning result of the target vehicle; determine the first element image based on the first positioning result and the high-precision map; perform a reliability test on the first positioning result based on the first element image to obtain the first detection result of the first positioning result.

[0118] It should be noted that the implementation of each operation can also be referred to accordingly. Figure 2 The corresponding description of the method embodiments shown.

[0119] exist Figure 7 In the described vehicle-mounted terminal 710, after acquiring the first positioning result of the target vehicle, a first element image is determined based on the first positioning result and a high-precision map. Then, a reliability test is performed on the first positioning result based on the first element image to obtain a first detection result of the first positioning result. Because the first positioning result is converted into a binary image based on the high-precision map, the first positioning result is compared with the high-precision map to detect whether the first positioning result is reliable. If the first positioning result is reliable, it can be used, effectively reducing the harm caused by unreliable positioning and helping to reduce safety hazards in autonomous driving.

[0120] This application also provides a chip, which includes at least one processor, a memory, and an interface circuit. The memory, the transceiver, and the at least one processor are interconnected via circuits. The at least one memory stores a computer program. When the computer program is executed by the processor... Figure 2 The method and flow shown are thus implemented.

[0121] This application also provides a computer-readable storage medium storing a computer program that, when run on an in-vehicle terminal,... Figure 2 The method and flow shown are thus implemented.

[0122] This application also provides a computer program product, which, when run on an in-vehicle terminal, provides a solution for... Figure 2 The method and flow shown are thus implemented.

[0123] This application also provides an intelligent vehicle, which includes a control system that performs the above-described actions. Figure 2 The method shown is used to control the driving of the aforementioned intelligent vehicle.

[0124] In summary, by implementing the embodiments of this application, after obtaining the first positioning result of the target vehicle, a first element image is determined based on the first positioning result and a high-precision map. Then, the reliability of the first positioning result is detected based on the first element image to obtain a first detection result of the first positioning result. Since the first positioning result is converted into a binary image based on the high-precision map, the first positioning result is compared with the high-precision map to detect whether the first positioning result is reliable. When the first positioning result is reliable, it is utilized, effectively reducing the harm caused by unreliable positioning and helping to reduce safety hazards in autonomous driving. In other words, this application can address scenarios where positioning is achieved using a positioning algorithm but lacks a true positioning value, providing real-time reliability information for the positioning result. This overcomes the shortcomings of existing technologies that cannot evaluate the reliability of positioning results without a true positioning value in real time. By using high-precision map projection, the positioning result is converted into a binary image and projected onto the high-precision map, allowing for comparison between the positioning result and the high-precision map to determine whether the positioning result is reliable. Utilizing reliable positioning results effectively reduces the harm caused by unreliable positioning results.

[0125] It should be understood that the processor mentioned in the embodiments of this application can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0126] It should also be understood that the memory mentioned in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0127] It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA, or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, the memory (storage module) is integrated into the processor.

[0128] It should be noted that the memories described herein are intended to include, but are not limited to, these and any other suitable types of memories.

[0129] It should also be understood that the first, second, third, fourth and various numerical designations used herein are merely for descriptive convenience and are not intended to limit the scope of this application.

[0130] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0131] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0132] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0133] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0134] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0135] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0136] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0137] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods shown in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0138] The steps in the method of this application embodiment can be adjusted, combined, or deleted according to actual needs.

[0139] The modules in the device of this application embodiment can be merged, divided, and deleted according to actual needs.

[0140] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for detecting positioning reliability, characterized in that, Applied to vehicle-mounted terminals, the method includes: Obtain the first location result of the target vehicle; A first element image is determined based on the first positioning result and the high-precision map. The first element image is an image in the forward-looking direction of the target vehicle. The reliability of the first positioning result is tested based on the first feature image to obtain a first detection result of the first positioning result; wherein, if the position and / or orientation of the map feature points in the first feature image matches the position and / or orientation of the map feature points seen when driving a vehicle normally, then the first positioning result is reliable.

2. The method according to claim 1, characterized in that, The step of performing reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result includes: The first element image is input into the target detection model for processing to obtain the first detection result of the first localization result; The target detection model is implemented based on a visual geometric group network architecture.

3. The method according to claim 2, characterized in that, Before inputting the first element image into the target detection model for processing to obtain the first detection result of the first localization result, the method further includes: Obtain N second positioning results and M third positioning results for the vehicle, wherein the second positioning results are historical positioning results with normal positioning, the third positioning results are historical positioning results with abnormal positioning, and N and M are positive integers; N second feature images are determined based on the N second positioning results and the high-precision map, and M third feature images are determined based on the M third positioning results and the high-precision map, wherein the N second positioning results correspond one-to-one with the N second feature images, and the M third positioning results correspond one-to-one with the M third feature images. The initial detection model is trained using the N second-element images and the M third-element images to obtain the target detection model; The initial detection model is implemented based on the visual geometry group network architecture.

4. The method according to claim 3, characterized in that, Based on the target location results and high-precision map, the target feature image is determined through the following steps: Map feature point data within a preset area are obtained from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points; The first coordinate of each of the plurality of map feature points is converted into a second coordinate, wherein the first coordinate of each map feature point is the coordinate of the map feature point in the world coordinate system, and the second coordinate is the coordinate of the map feature point in the second coordinate system, wherein the second coordinate system is a coordinate system with the positioning coordinate in the target positioning result as the origin. The second coordinates of each map feature point are converted into image coordinates based on the camera intrinsic parameters of the vehicle terminal to obtain the target feature image, wherein the pixel coordinates of the target feature image are the image coordinates; Wherein, if the target positioning result is the first positioning result, then the target element image is the first element image; If the target location result is the second location result, then the target element image is the second element image; If the target positioning result is the third positioning result, then the target element image is the third element image.

5. The method according to claim 4, characterized in that, The first coordinate is converted to the second coordinate according to the following formula: Among them, (X) w ,Y w Z w (X) is the first coordinate, (X) c ,Y c Z c R is the second coordinate, R is the rotation matrix of the three degrees of freedom of the target positioning result, and T is the translation matrix of the positioning coordinate in the target positioning result.

6. The method according to claim 5, characterized in that, The second coordinate is converted into image coordinates according to the following formula: Where (u,v) are the image coordinates, f x f y Let u0 and v0 be the focal length of the camera in the vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

7. The method according to any one of claims 1-6, characterized in that, The method further includes: If the first positioning result comes directly from the positioning sensor, then the positioning error covariance determined by the positioning sensor for the first positioning result is obtained; A second detection result is determined based on the first detection result and the positioning error covariance.

8. The method according to claim 7, characterized in that, The method further includes: If the first detection result and the second detection result indicate that the first positioning result is normal, then the first positioning result is used for autonomous driving; If the first detection result and / or the second detection result indicate that the first positioning result is abnormal, a warning will be issued to disable autonomous driving.

9. A positioning reliability detection device, characterized in that, The device, applied to an in-vehicle terminal, includes: The acquisition unit is used to acquire the first positioning result of the target vehicle. The determining unit is configured to determine a first element image based on the first positioning result and the high-precision map, wherein the first element image is an image in the forward-looking direction of the target vehicle; The detection unit is configured to perform reliability detection on the first positioning result based on the first element image to obtain a first detection result of the first positioning result; wherein, if the position and / or orientation of the map element points in the first element image matches the position and / or orientation of the map element points seen when driving a vehicle normally, the first positioning result is reliable.

10. The apparatus according to claim 9, characterized in that, The detection unit is used for: The first element image is input into the target detection model for processing to obtain the first detection result of the first localization result; The target detection model is implemented based on a visual geometric group network architecture.

11. The apparatus according to claim 10, characterized in that, The device further includes a training unit for: Obtain N second positioning results and M third positioning results for the vehicle, wherein the second positioning results are historical positioning results with normal positioning, the third positioning results are historical positioning results with abnormal positioning, and N and M are positive integers; N second feature images are determined based on the N second positioning results and the high-precision map, and M third feature images are determined based on the M third positioning results and the high-precision map, wherein the N second positioning results correspond one-to-one with the N second feature images, and the M third positioning results correspond one-to-one with the M third feature images. The initial detection model is trained using the N second-element images and the M third-element images to obtain the target detection model; The initial detection model is implemented based on the visual geometry group network architecture.

12. The apparatus according to claim 11, characterized in that, The determining unit is used for: Map feature point data within a preset area are obtained from the high-precision map, wherein the preset area is determined based on the target positioning result, and the map feature point data includes multiple map feature points; The first coordinate of each of the plurality of map feature points is converted into a second coordinate, wherein the first coordinate of each map feature point is the coordinate of the map feature point in the world coordinate system, and the second coordinate is the coordinate of the map feature point in the second coordinate system, wherein the second coordinate system is a coordinate system with the positioning coordinate in the target positioning result as the origin. The second coordinates of each map feature point are converted into image coordinates based on the camera intrinsic parameters of the vehicle terminal to obtain the target feature image, wherein the pixel coordinates of the target feature image are the image coordinates; Wherein, if the target positioning result is the first positioning result, then the target element image is the first element image; If the target location result is the second location result, then the target element image is the second element image; If the target positioning result is the third positioning result, then the target element image is the third element image.

13. The apparatus according to claim 12, characterized in that, The first coordinate is converted to the second coordinate according to the following formula: Among them, (X) w ,Y w Z w (X) is the first coordinate, (X) c ,Y c Z c R is the second coordinate, R is the rotation matrix of the three degrees of freedom of the target positioning result, and T is the translation matrix of the positioning coordinate in the target positioning result.

14. The apparatus according to claim 13, characterized in that, The second coordinate is converted into image coordinates according to the following formula: Where (u,v) are the image coordinates, f x f y Let u0 and v0 be the focal length of the camera in the vehicle terminal, and u0 and v0 be the principal point coordinates relative to the imaging plane.

15. The apparatus according to any one of claims 9-14, characterized in that, The detection unit is also used for: If the first positioning result comes directly from the positioning sensor, then the positioning error covariance determined by the positioning sensor for the first positioning result is obtained; A second detection result is determined based on the first detection result and the positioning error covariance.

16. The apparatus according to claim 15, characterized in that, The device further includes a processing unit for: If the first detection result and the second detection result indicate that the first positioning result is normal, then the first positioning result is used for autonomous driving; If the first detection result and / or the second detection result indicate that the first positioning result is abnormal, a warning will be issued to disable autonomous driving.

17. A vehicle-mounted terminal, characterized in that, The method includes a processor, a memory, a communication interface, and one or more programs, said one or more programs being stored in the memory and configured to be executed by the processor, said programs including instructions for performing the steps of the method as described in any one of claims 1-8.

18. A chip, characterized in that, include: A processor for retrieving and running a computer program from memory, causing a device on which the chip is mounted to perform the method as described in any one of claims 1-8.

19. A computer-readable storage medium, characterized in that, It stores a computer program for electronic data interchange, wherein the computer program causes the computer to perform the method as described in any one of claims 1-8.

20. A computer program product comprising a computer program that, when run on a computer, implements the method as described in any one of claims 1-8.

21. An intelligent vehicle, characterized in that, The intelligent vehicle includes a control system that performs the method as described in any one of claims 1-8 to control the driving of the intelligent vehicle.

Images