Positioning error detection method, device and storage medium

Abstract

The application relates to a positioning error detection method and device and a storage medium. The positioning error detection device obtains positioning data and environment data of a vehicle. The environment data represents image data of the surrounding environment of the vehicle during driving. Road marking information is determined based on the environment data and the positioning data. The road marking information represents information of road markings corresponding to the driving track of the vehicle. Feature information to be processed is determined according to the road marking information, and a positioning error detection result of the vehicle is determined based on the feature information to be processed and a trained positioning error detection model. Thus, effective detection of the positioning error can be realized, thereby improving the accuracy of positioning.

Description

Technical Field

[0001] This application relates to the field of positioning technology, and in particular to a positioning error detection method, device and storage medium. Background Technology

[0002] In the related technologies of intelligent driving, massive amounts of map element data are mainly collected in a crowdsourced manner, and maps are created through post-processing work such as map learning. However, although this method can create maps, the creation logic is simple and there will be positioning errors. Therefore, how to effectively detect positioning errors in map positioning is a technical problem that urgently needs to be solved. Summary of the Invention

[0003] This application provides a positioning error detection method, device, and storage medium, which can effectively detect positioning errors and thus improve positioning accuracy.

[0004] In a first aspect, embodiments of this application provide a positioning error detection method, the method comprising:

[0005] Acquire vehicle location data and environmental data; among which, environmental data represents image data of the vehicle's surrounding environment during driving.

[0006] Road marking information is determined based on environmental and location data; the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0007] The feature information to be processed is determined based on the road marking information, and the positioning error detection result of the vehicle is determined based on the feature information to be processed and the trained positioning error detection model.

[0008] In this embodiment, the positioning error detection device can receive the vehicle's own positioning data and the image data of its surrounding environment in real time during the vehicle's operation. Then, it can use the positioning data and environmental data to determine the road marking information corresponding to the vehicle's driving trajectory, i.e., the information of the road markings along the route during the driving process. Based on this road marking information, it can extract the feature information to be processed, and use the feature information to be processed and the trained positioning error detection model to determine the positioning error detection result, thereby achieving effective detection of the vehicle's positioning error and improving the accuracy of vehicle positioning.

[0009] Furthermore, in some embodiments, determining road marking information based on environmental data and location data includes:

[0010] The vehicle's field of view area is determined based on the positioning data; where the field of view area represents the area of ​​the vehicle's field of view during the driving process.

[0011] Road marking information in the field of view area is determined based on the spatial location information of the environmental data.

[0012] In this embodiment, the positioning error detection device can first use positioning data to determine the field of view area of ​​the vehicle during driving. Then, based on the spatial location information of the environmental data, it can determine the road marking information falling within the field of view area. This allows it to identify road markings related to the vehicle's driving process. By using these road markings to perform subsequent positioning error detection, the accuracy of positioning error detection can be effectively improved.

[0013] Furthermore, in some embodiments, determining the vehicle's field of view region based on positioning data includes:

[0014] The vehicle's direction of travel is determined based on the location data;

[0015] The coordinates of the field of view boundary points are determined based on the location information of the vehicle's cameras and its driving direction.

[0016] The field of view area is determined based on the location information of the vehicle's camera and the coordinate information of the field of view boundary point.

[0017] In this embodiment, when determining the field of view area, the positioning error detection device can first determine the vehicle's driving direction based on the positioning data, and then use the vehicle's camera as a reference to determine the coordinate information of the field of view boundary point in combination with the driving direction. Thus, the field of view area can be determined using the camera's position information and the coordinate information of the field of view boundary point. This allows for the clear identification of the field of view area during the vehicle's movement, thereby enabling accurate positioning error detection using the field of view area.

[0018] Furthermore, in some embodiments, determining the feature information to be processed based on road marking information includes:

[0019] Obtain the combined navigation reference information corresponding to the road marking information, and determine the first feature information based on the combined navigation reference information; wherein, the combined navigation reference information includes at least one of the solution status information, satellite search number information, and speed information output by the vehicle's combined navigation module;

[0020] The second feature information is determined based on road marking information and camera location information;

[0021] The third feature information is determined based on road marking information and location data;

[0022] The feature information to be processed is determined based on the first feature information, the second feature information, and the third feature information.

[0023] In this embodiment, the feature information to be processed may include features of different dimensions or types, including first feature information determined by the combined navigation reference information output by the vehicle's combined navigation module, second feature information determined by road marking information and camera position information, and third feature information determined by road marking information and positioning data. Thus, positioning error detection can be completed based on the input of multi-dimensional feature information into the trained positioning error detection model, thereby improving the effect of positioning error detection.

[0024] Furthermore, in some embodiments, before determining the feature information to be processed based on road marking information, and determining the vehicle's positioning error detection result based on the feature information to be processed and the trained positioning error detection model, the method further includes:

[0025] Obtain the first road data;

[0026] The first road data is filtered to obtain the second road data; wherein the data filtering process includes at least one of the filtering process of road mark shape and the filtering process of data attributes.

[0027] The initial positioning error detection model is trained based on the second road data and the ground truth map data to obtain the trained positioning error detection model.

[0028] In this embodiment, the positioning error detection device can filter out data in the first road data that does not meet the corresponding conditions by filtering road mark patterns and / or data attributes. This can fully take into account the differences and influences of various types of road data, improve the consistency of the data used to train the positioning error detection model, and then use the filtered second road data and ground truth map data to train the initial positioning error detection model, which can effectively improve the positioning error detection effect of the trained positioning error detection model.

[0029] Furthermore, in some embodiments, the first road data is subjected to data filtering processing to obtain second road data, including:

[0030] Determine the bounding rectangle of the road marker to be processed in the first road data;

[0031] Based on the first morphological information of the circumscribed rectangle and the second morphological information of the road markers to be processed in the ground truth map data, the road marker morphology is filtered to obtain the second road data.

[0032] In this embodiment, when the positioning error detection device performs road mark morphology filtering, it can first determine the circumscribed rectangle of the road mark to be processed in the first road data, that is, determine the minimum circumscribed rectangle of the road mark to be processed. Then, it can use the circumscribed rectangle to determine the first morphology information of the road mark to be processed. Then, it can use the first morphology information and the second morphology information corresponding to the road mark to be processed in the ground truth map data, that is, the true morphology information of the road mark to be processed, to filter the road mark to be processed. In this way, the road data can be filtered according to the road mark morphology, thereby improving the data consistency and effectiveness of the second road data.

[0033] Furthermore, in some embodiments, the first road data is subjected to data filtering processing to obtain second road data, including:

[0034] Based on the category attribute identification information of the first road data, identify the identification information to be processed in the first road data that does not meet the preset identification conditions;

[0035] The road data corresponding to the identification information to be processed in the first road data is removed to complete the data attribute filtering process and obtain the second road data.

[0036] In this embodiment, when the positioning error detection device performs data attribute filtering, it can remove road data corresponding to the unprocessed identifier information that does not meet the preset identifier conditions in the category attribute identifier information of the first road data. This ensures that the category attributes of the filtered road data all meet the requirements, thereby obtaining second road data that meets the category attribute requirements, improving the data consistency of the second road data, and thus improving the training effect of the positioning error detection model.

[0037] Secondly, embodiments of this application provide a positioning error detection device, which includes an acquisition unit, a determination unit, and a positioning error detection unit;

[0038] The acquisition unit is used to acquire the vehicle's positioning data and environmental data; wherein, the environmental data represents image data of the vehicle's surrounding environment during the driving process.

[0039] The determining unit is used to determine road marking information based on environmental data and positioning data; wherein, the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory;

[0040] The positioning error detection unit is used to determine the feature information to be processed based on the road marking information, and to determine the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model.

[0041] Thirdly, embodiments of this application provide a positioning error detection device, which includes a processor and a memory storing processor-executable instructions; when the instructions are executed by the processor, the above-mentioned positioning error detection method is implemented.

[0042] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described positioning error detection method. Attached Figure Description

[0043] Figure 1 Schematic diagram of the implementation process of the positioning error detection method provided in the embodiments of this application Figure 1 ;

[0044] Figure 2 Schematic diagram of the implementation process of the positioning error detection method provided in the embodiments of this application Figure 2 ;

[0045] Figure 3 Schematic diagram of the implementation process of the positioning error detection method provided in the embodiments of this application Figure 3 ;

[0046] Figure 4 Schematic diagram of the implementation of the positioning error detection method provided in the embodiments of this application Figure 1 ;

[0047] Figure 5 Schematic diagram of the implementation of the positioning error detection method provided in the embodiments of this application Figure 2 ;

[0048] Figure 6 Schematic diagram of the composition structure of the positioning error detection device provided in the embodiments of this application Figure 1 ;

[0049] Figure 7 Schematic diagram of the composition structure of the positioning error detection device provided in the embodiments of this application Figure 2 . Detailed Implementation

[0050] The embodiments of this application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be understood that the preferred embodiments are only for illustrating this application and are not intended to limit the scope of protection of this application.

[0051] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0052] In the related technologies of intelligent driving, massive amounts of map element data are mainly collected in a crowdsourced manner, and maps are created through post-processing work such as map learning. However, although this method can create maps, the creation logic is simple and there will be positioning errors. Therefore, how to effectively detect positioning errors in map positioning is a technical problem that urgently needs to be solved.

[0053] To address the aforementioned issues, this application proposes a positioning error detection method, apparatus, and storage medium. The positioning error detection apparatus acquires vehicle positioning data and environmental data; wherein the environmental data represents image data of the vehicle's surrounding environment during driving; road marking information is determined based on the environmental data and positioning data; wherein the road marking information represents information about road markings corresponding to the vehicle's driving trajectory; feature information to be processed is determined based on the road marking information, and the positioning error detection result of the vehicle is determined based on the feature information to be processed and the trained positioning error detection model; based on the above scheme, effective detection of positioning errors can be achieved, thereby improving positioning accuracy.

[0054] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0055] Figure 1 Schematic diagram of the implementation process of the positioning error detection method provided in the embodiments of this application Figure 1 ,like Figure 1 As shown in the embodiments of this application, the positioning error detection method of the positioning error detection device may include the following steps:

[0056] Step 101: Obtain vehicle positioning data and environmental data; wherein, the environmental data represents image data of the vehicle's surrounding environment during the driving process.

[0057] In the embodiments of this application, the positioning error detection device can acquire vehicle positioning data and environmental data; wherein, the environmental data represents image data of the vehicle's surrounding environment during driving.

[0058] It should be noted that, in the embodiments of this application, the vehicle can acquire environmental data of its surrounding environment in real time during movement. The vehicle can acquire environmental data through various built-in sensor devices. For example, the vehicle can take pictures of the real-time environment through its built-in visual sensors, such as cameras, and can also detect the surrounding environment through radar to obtain environmental data.

[0059] For example, environmental data may include image data of the road surface on which the vehicle travels, as well as image data of other vehicles on the road, and may also include image data of buildings or plants on both sides of the road.

[0060] Understandably, vehicles can generate real-time location data, which is data used to indicate their real-time location.

[0061] Step 102: Determine road marking information based on environmental data and location data; wherein, road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0062] In the embodiments of this application, the positioning error detection device can determine road marking information based on the vehicle's positioning data and environmental data after acquiring the vehicle's positioning data and environmental data; wherein, the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0063] In some embodiments of this application, the specific road markings included in the road marking information are not limited thereto; for example, the road marking information may include markings such as arrows, lane lines, and road boundaries on the road.

[0064] In some embodiments of this application, when determining road marking information based on environmental data and positioning data, the positioning error detection device can determine the field of view area of ​​the vehicle based on the positioning data; and then determine the road marking information in the field of view area based on the spatial location information of the environmental data.

[0065] In the embodiments of this application, the field of view region represents the region of the vehicle's field of view during driving.

[0066] In some embodiments of this application, the field of view area can be the field of view area of ​​the vehicle's front-facing camera, that is, the area corresponding to the physical angle of the image range that the vehicle's front-facing camera can capture; for example, the field of view area of ​​the vehicle's front-facing camera can be used, wherein the field of view of the front-facing camera is 60 degrees by default, and the field of view area can be the area range corresponding to the 60-degree field of view of the front-facing camera in the default state. However, during the driving process of the vehicle, the field of view may change, and the field of view area can change accordingly.

[0067] In some embodiments of this application, when the positioning error detection device determines the field of view area of ​​a vehicle based on positioning data, it can determine the driving direction of the vehicle based on the positioning data; then determine the coordinate information of the field of view boundary point based on the position information of the vehicle's camera and the driving direction; thereby determining the field of view area based on the position information of the vehicle's camera and the coordinate information of the field of view boundary point.

[0068] In some embodiments of this application, the vehicle may have multiple cameras, and this application does not make specific limitations; for example, if the vehicle's camera is a front-facing camera, then the location information of the vehicle's cameras can be the location information of the front-facing camera.

[0069] In some embodiments of this application, when determining the field of view region, the vehicle's camera can be used as the vertex of the field of view region to determine the field of view region.

[0070] In some embodiments of this application, the field of view region is constructed using the vehicle's front-facing camera. The vehicle's driving direction can be determined first based on positioning data, and the vehicle's driving direction can be used as the center direction of the front-facing camera's field of view. The range of the field of view region can be defined based on the left and right boundary lines of the field of view. Therefore, the coordinate information of the field of view boundary points can be determined, and the left and right boundary lines of the field of view can be determined based on the coordinate information of the field of view boundary points, thereby determining the field of view region.

[0071] In some embodiments of this application, the field-of-view boundary point coordinate information may include the coordinates of the left boundary point and the coordinates of the right boundary point; the camera position information may be represented in the form of coordinates; when determining the field-of-view boundary point coordinate information based on the vehicle's camera position information and driving direction, the current field-of-view angle of the camera may be determined based on the vehicle's driving direction, and the current field-of-view angle represents the field-of-view angle of the vehicle's camera based on the dynamic driving direction during the vehicle's driving process; wherein, the current field-of-view angle may include a first angle and a second angle; the first angle represents the left angle in the current field-of-view angle, the second angle represents the right angle in the current field-of-view angle, and the addition of the left angle and the right angle yields the current field-of-view angle; then, the left boundary point coordinates may be determined based on the first angle, the coordinates of the camera's position information, and the observation distance, and the right boundary point coordinates may be determined based on the second angle, the coordinates of the camera's position information, and the observation distance; wherein, the observation distance represents the maximum distance of the camera in the field of view direction of the captured image.

[0072] For example, assuming the observation distance d = 10 meters, the coordinates of the vehicle's front-facing camera's position information are p(x,y), and the first field of view is denoted as direction, then the coordinates of the left boundary point can be determined by the following formula:

[0073] P left=(x+d·cos(left) angle ),y+d·sin(left angle (1)

[0074] Where, left angle That is, the first angle, left. angle It can be determined using the following formula:

[0075]

[0076] Here, angle represents the default field of view, for example, angle can be 60 degrees.

[0077] The coordinates of the right boundary point can be determined using the following formula:

[0078] P right =(x+d·cos(right) angle ),y+d·sin(right angle (3)

[0079] Among them, right angle That is, the second angle, right. angle It can be determined using the following formula:

[0080]

[0081] In some embodiments of this application, when determining the field of view region based on the position information of the vehicle's camera and the coordinate information of the field of view boundary point, the coordinates of the vehicle's front camera can be used as the vertex of the field of view, which together with the coordinates of the left boundary point and the right boundary point in the field of view boundary point coordinate information forms a triangular region, which is the field of view region.

[0082] For example, the geometric object of the field of view region can be generated based on the polygon constructor, the coordinates of the front-facing camera, the coordinates of the left boundary point, and the coordinates of the right boundary point. This can be expressed by the following formula:

[0083] view cone =Polygon[P,P left ,P right (5)

[0084] In some embodiments of this application, when determining road marking information in the field of view region based on the spatial location information of environmental data, candidate markings in the environmental data can be determined first, and then intersection operations can be performed based on the spatial location information of the candidate markings and the field of view region to determine the road marking information in the field of view region; wherein, the candidate markings can be markings on the road in the environmental data, such as lane lines, arrows and road boundary markings.

[0085] It should be noted that road marking information represents information about road markings that fall within the field of view area.

[0086] In some embodiments of this application, the minimum bounding rectangle of each candidate mark can be determined based on the spatial location information of the environmental data, and then the geometric center of the minimum bounding rectangle of the candidate mark can be used as the position coordinate of the candidate mark, thereby determining the road mark information based on the candidate mark corresponding to the position coordinate in the field of view area.

[0087] For example, the intersection operation can be performed on the spatial location information of the field of view area and the candidate mark. If the geometric center of the smallest bounding rectangle of a candidate mark falls within the range of the field of view area, it can be determined that the candidate mark is related to the vehicle's driving path and can be used as a road mark. This determines the road mark information, which can quickly filter out road marks within the field of view area, improve matching efficiency, and reduce interference from other irrelevant elements in the road.

[0088] Step 103: Determine the feature information to be processed based on the road marking information, and determine the vehicle's positioning error detection result based on the feature information to be processed and the trained positioning error detection model.

[0089] In the embodiments of this application, the positioning error detection device can determine the feature information to be processed based on the road marking information after determining the road marking information based on the environmental data and positioning data, and determine the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model.

[0090] In the embodiments of this application, the feature information to be processed may include features of multiple dimensions or types, thereby enriching the expression of the feature information to be processed and improving the effect of positioning error detection.

[0091] It is understood that in the embodiments of this application, the trained positioning error detection model is a model obtained through training. This application uses a Long Short-Term Memory (LSTM) model as the positioning error detection model for training, thereby obtaining the trained positioning error detection model.

[0092] In some embodiments of this application, the encoder in the trained localization error detection model can map the input feature information to be processed to a high-dimensional space and extract multi-dimensional features therein. The decoder in the trained localization error detection model can then use the multi-dimensional features obtained from the encoder to predict the localization error and obtain the localization error detection result.

[0093] In some embodiments of this application, the temporal dependency forget gate mechanism in the trained positioning error detection model can capture long-term dependencies and nonlinear dynamics, namely the temporal effects of the distribution of absolute position errors of map features during continuous acquisition.

[0094] In some embodiments of this application, when the positioning error detection device determines the feature information to be processed based on the road marking information, it can obtain the combined navigation reference information corresponding to the road marking information and determine the first feature information based on the combined navigation reference information; wherein, the combined navigation reference information includes at least one of the solution status information, satellite search number information, and speed information output by the vehicle's combined navigation module; then, it determines the second feature information based on the road marking information and the camera's position information; further, it determines the third feature information based on the road marking information and positioning data; thereby determining the feature information to be processed based on the first feature information, the second feature information, and the third feature information.

[0095] In some embodiments of this application, the vehicle may be equipped with a combined navigation module, which is an integrated system that combines multiple positioning, path planning, environmental perception and other technologies; the combined navigation module can generate combined navigation reference information.

[0096] In some embodiments of this application, the integrated navigation module can generate and output the solution status, satellite search count, and speed information of the trajectory points corresponding to the road marking information. After receiving this information, the positioning error detection device can fuse this information into the feature vector to determine the first feature information.

[0097] In some embodiments of this application, the first feature information may include a first component corresponding to the solution's state information, a second component corresponding to the number of satellites searched, and a third component corresponding to the speed information; wherein, the first component can be determined based on the trajectory points corresponding to the solution's state information and road marking information; the second component can be determined based on the trajectory points corresponding to the number of satellites searched and road marking information; and the third component can be determined based on the trajectory points corresponding to the speed information and road marking information.

[0098] For example, the trajectory points corresponding to the road marking information are represented as {c1, c2, ... c n Then the integrated navigation module can output the state of the solution for these trajectory points. Search star number and speed information The positioning error detection device acquires this integrated navigation reference information and determines, based on this information, a first component corresponding to the solution's state information, a second component corresponding to the number of satellites searched, and a third component corresponding to the velocity information. The first component can be determined using the following formula (6), the second component using the following formula (7), and the third component using the following formula (8).

[0099]

[0100] Where n represents the total number of trajectory points corresponding to the road marking information; Represents the state of the solution The frequency function, the third component can be regarded as The mode.

[0101] In the embodiments of this application, the reliability of the solution state can reflect the positioning accuracy of the vehicle, the number of satellite searches can reflect the availability and stability of satellite signals, and speed information can be used as a supplementary verification index for the dynamic state of the vehicle.

[0102] In some embodiments of this application, the position information of the vehicle's camera can be regarded as the position of the vehicle, so that the position information of the vehicle's camera can be modeled as the vehicle's driving trajectory; the camera on the vehicle moves along the lane line, and it is assumed that the camera's direction of travel is consistent with the vehicle's direction of travel. The shortest distance between the centroid of each road mark in the road mark information and the position information of the vehicle's camera can be extracted. This distance can be regarded as the mutual positional relationship between the road mark and the vehicle position, and this distance is used as the second feature information; the second feature information can reflect the spatial relationship between the vehicle and the road mark.

[0103] In some embodiments of this application, when determining the third feature information based on road marking information and positioning data, a matching mark is generated based on the positioning data to be matched with the road markings in the road marking information; then the geometric centroid of each road marking in the road marking information is determined, the Euclidean distance between the geometric centroid of the road marking and the matching mark is calculated, and the shortest distance in the Euclidean distance calculation result is used as the third feature information; the positional error of the matching road marking can be reflected through the third feature information.

[0104] For example, the feature information to be processed can be represented as λ = (λ1, λ2, λ3, λ4, λ5), where λ1 is the first component, λ2 is the second component, λ3 is the third component, λ4 is the second feature information, and λ5 is the third feature information; thus, feature information can be constructed from various dimensions, including spatial relationships and positional relationships, to improve the accuracy of positioning error detection.

[0105] In some embodiments of this application, the positioning error detection device can predict the positioning error of the vehicle at future times based on the currently received data. In other words, the positioning error detection result can be the result of the positioning error detection device predicting the positioning error of the vehicle at future times.

[0106] In some embodiments of this application, such as Figure 2 As shown, before the positioning error detection device determines the feature information to be processed based on road marking information, and determines the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model, it may further include the following steps:

[0107] Step 104: Obtain the first road data.

[0108] In the embodiments of this application, before the positioning error detection device determines the feature information to be processed based on the road marking information and determines the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model, it can acquire the first road data.

[0109] In some embodiments of this application, the first road data may be different road data collected by the vehicle through Simultaneous Localization and Mapping (SLAM). The first road data may include image data of the road surface and image data of other parts of the road surface, such as buildings or plants.

[0110] In some embodiments of this application, the first road data can be preprocessed, including encoding various types of data in the first road data to obtain category attribute identifiers for each type of data.

[0111] Step 105: Perform data filtering processing on the first road data to obtain the second road data; wherein, the data filtering processing includes at least one of road mark pattern filtering processing and data attribute filtering processing.

[0112] In the embodiments of this application, after acquiring the first road data, the positioning error detection device can perform data filtering processing on the first road data to obtain the second road data; wherein, the data filtering processing includes at least one of road mark pattern filtering processing and data attribute filtering processing.

[0113] In the embodiments of this application, data filtering processing can be understood as removing low-quality or invalid data from the first road data, thereby ensuring the accuracy and completeness of the obtained second road data and improving the training effect of the positioning error detection model.

[0114] In the embodiments of this application, the road mark pattern filtering process represents the process of removing data in the first road data that does not conform to the road mark pattern; the data attribute filtering process represents the process of removing data in the first road data that does not conform to the data attribute requirements.

[0115] In some embodiments of this application, when the positioning error detection device performs data filtering processing on the first road data to obtain the second road data, it can determine the circumscribed rectangle of the road mark to be processed in the first road data; then, based on the first morphological information of the circumscribed rectangle and the second morphological information corresponding to the road mark to be processed in the ground truth map data, it performs road mark morphological filtering processing to obtain the second road data.

[0116] In embodiments of this application, the road markings to be processed may include planar road markings such as lane lines and arrows on the road surface.

[0117] In the embodiments of this application, the circumscribed rectangle represents the minimum circumscribed rectangle of the road mark to be processed; the circumscribed rectangle of the road mark to be processed in the first road data can be extracted by the minimum circumscribed rectangle (MBR) algorithm; wherein, the MBR algorithm is implemented based on the rotating caliper algorithm, which finds the circumscribed rectangle with the smallest area by traversing the rotation angle of the polygon vertices.

[0118] In some embodiments of this application, after obtaining the circumscribed rectangle of the road mark to be processed, the first morphological information of the circumscribed rectangle can be determined. When determining the first morphological information of the circumscribed rectangle, the four vertices of the circumscribed rectangle can be determined first, and then the lengths of the four sides of the circumscribed rectangle can be determined based on the four vertices. Then, the lengths of the four sides are sorted according to their lengths, so as to determine the first morphological information based on the sorted side lengths.

[0119] In some embodiments of this application, when sorting the side lengths, the longest side can be defined as the long side, and the remaining second longest side can be defined as the short side.

[0120] In some embodiments of this application, the first morphological information may include side length information and included angle information; wherein, the side length information may include the length information of the side length of the circumscribed rectangle, and the included angle information includes the angle information of the included angle formed by the side lengths of the circumscribed rectangle.

[0121] It is understandable that true value map data can be used as reference map data, that is, it can be used as real map data for reference.

[0122] In some embodiments of this application, the second morphological information may include the side length vector data and included angle data of the road mark to be processed in the truth map data.

[0123] For example, if the coordinates of the four vertices of the bounding rectangle of a road marking, such as an arrow, are P1(x1,y1), P2(x2,y2), P3(x3,y3), and P4(x4,y4), then the vector of the longer side can be represented by the following formula:

[0124]

[0125] L1 and L2 can be represented by the following formulas:

[0126]

[0127] The included angle information can be determined based on the magnitudes of the vectors of the longer and shorter sides. For example, the included angle information can be expressed by the following formula:

[0128]

[0129] Among them, V 参考 This represents the side length vector corresponding to the road marker to be processed in the true map data.

[0130] For example, based on the four vertices of the circumscribed rectangle of the arrow in the above example, the vector of the shorter side can be represented by the following formula:

[0131]

[0132] In some embodiments of this application, if the calculated included angle information does not conform to a preset included angle threshold range, it can be determined that the road mark to be processed in the first road data will be removed to obtain the second road data. The value of the preset included angle threshold range is not specifically limited in this application. For example, the preset included angle threshold range can be determined by a first included angle threshold and a second included angle threshold. The first included angle threshold can be less than the second included angle threshold. If the included angle information is less than the first included angle threshold or greater than the second included angle threshold, it can be determined that it does not conform to the preset included angle threshold range, meaning the road mark to be processed does not conform to the shape of a road mark and needs to be removed. The preset included angle threshold range can also be determined by a third included angle threshold. For example, the preset included angle threshold range can be a range less than the third included angle threshold. When the included angle information is greater than or equal to the third included angle threshold, it can be determined that it does not conform to the shape of a road mark and the road mark to be processed will be removed.

[0133] In some embodiments of this application, when performing road mark shape screening, the length ratio of the long side to the short side can also be determined; when the ratio of the long side to the short side does not meet the preset ratio threshold range, and the modulus of the long side or the modulus of the short side does not meet their respective modulus threshold range, it can be determined that the road mark to be processed does not meet the shape of a road mark and needs to be removed.

[0134] In some embodiments of this application, the preset ratio threshold range is not specifically limited; for example, the preset ratio range can be less than 3 and greater than or equal to 1.2.

[0135] In some embodiments of this application, the range of the modulus threshold is not specifically limited; for example, the modulus threshold range of the long side can be less than or equal to 2.5, and the modulus threshold range of the short side can be greater than or equal to 0.3.

[0136] For example, the ratio of the longer side to the shorter side can be expressed as R = V 长 / V 短 This allows us to compare R with a preset ratio threshold range. If R does not meet the preset ratio threshold range, we can continue to compare the modulus and modulus threshold range of the long side and the short side respectively. If the modulus of the long side does not meet the modulus threshold range of the long side, we can determine to remove the road mark to be processed.

[0137] In some embodiments of this application, when the positioning error detection device performs data filtering processing on the first road data to obtain the second road data, it can determine the unprocessed identification information in the first road data that does not meet the preset identification conditions based on the category attribute identification information of the first road data; then, it removes the road data corresponding to the unprocessed identification information in the first road data to complete the data attribute filtering processing and obtain the second road data.

[0138] In the embodiments of this application, the category attribute identification information can be used to indicate the category attributes of various types of data in the first road data. Different types of data can have different category attribute identifiers, and the category attribute identifiers of data of the same type have at least some of the same numerical representations. For example, the category attribute identifier of road markings in the first road data can be 1234XXXX, the category attribute identifier of non-road markings, such as buildings, can be 2222XXXX, and the attribute identifier of roadside plants can be 3333XXXX.

[0139] In the embodiments of this application, the preset identification condition may be that the category attribute identification information needs to include a numerical representation indicating the category to which the road mark belongs; for example, if the category attribute identification of the road mark is 1234XXXX, then the preset identification condition may be that the category attribute identification information includes a numerical representation with a four-digit value of "1234" from the left.

[0140] For example, if the preset identification condition is that the category attribute identification information includes a four-digit value of "1234" from the left, then for the first road data, the building category data with category attribute identification of 2222XXXX (identification information to be processed) can be directly removed, and the plant category data with category attribute identification of 3333XXXX (identification information to be processed) can also be directly removed. This enables filtering based on category attribute identification information and improves the filtering efficiency of data filtering.

[0141] Step 106: Train the initial positioning error detection model based on the second road data and the ground truth map data to obtain the trained positioning error detection model.

[0142] In the embodiments of this application, after the positioning error detection device performs data filtering processing on the first road data to obtain the second road data, it can train the initial positioning error detection model based on the second road data and the ground truth map data to obtain the trained positioning error detection model.

[0143] In some embodiments of this application, after obtaining the second road data, the road surface features in the second road data, such as arrows and lane lines on the road surface, can be used to calculate the position error with the ground truth features in the ground truth map data corresponding to the road surface features. That is, by calculating the Euclidean distance between the geometric centroid of the road surface feature and the centroid of the ground truth feature, the shortest distance is selected as the position error, and these position errors are used as a dataset for model training and validation.

[0144] For example, for arrows on the road surface, the Euclidean distance between the geometric centroid of the arrow feature in the second road data and the geometric centroid of the corresponding arrow ground truth feature in the ground truth map data can be calculated. Spatial indexing can be used to accelerate the matching process. The shortest distance among all calculated Euclidean distances is taken as the position error of the arrow feature, and the matching result between the arrow feature in the second road data and the corresponding arrow ground truth feature is recorded and saved.

[0145] For example, for lane lines on the road surface, the Euclidean distance between the geometric centroid of the lane line in the second road data and the geometric centroid of the corresponding lane line feature in the ground truth map data can be calculated. Similarly, spatial indexing can be used to accelerate the matching process. The shortest distance among all calculated Euclidean distances is taken as the position error of the lane line feature, and the matching result between the lane line feature in the second road data and the corresponding ground truth lane line feature is recorded and saved.

[0146] In some embodiments of this application, after training the positioning error detection model is completed, a validation set can be used to verify the effectiveness of the trained positioning error detection model in detecting positioning errors.

[0147] This application provides a positioning error detection method. The positioning error detection device can acquire vehicle positioning data and environmental data. The environmental data represents image data of the vehicle's surrounding environment during driving. Road marking information is determined based on the environmental data and positioning data. This road marking information represents information about road markings corresponding to the vehicle's driving trajectory. To-be-processed feature information is determined based on the road marking information, and the positioning error detection result is determined based on the to-be-processed feature information and a trained positioning error detection model. Therefore, the positioning error detection device can receive the vehicle's own positioning data and the image data of its surrounding environment in real time during vehicle operation. It can then use the positioning data and environmental data to determine the road marking information corresponding to the vehicle's driving trajectory, i.e., the information of road markings encountered during driving. Based on this road marking information, to-be-processed feature information can be extracted, and the positioning error detection result can be determined using the to-be-processed feature information and a trained positioning error detection model. This achieves effective detection of vehicle positioning errors, thereby improving the accuracy of vehicle positioning.

[0148] Based on the above embodiments, in another embodiment of this application, a positioning error detection method is provided, as exemplarily. Figure 3 As shown, the first road data can be preprocessed first (step 201). The preprocessing includes encoding various types of data in the first road data to obtain category attribute identifiers for each type of data. Then, the first road data is filtered to obtain the second road data (step 202). The data filtering process can include filtering based on road mark morphology and filtering based on data attributes. Then, the position error can be calculated based on the road surface features in the second road data and the ground truth features in the ground truth map data to construct a training set and a validation set (step 203). Next, the positioning error detection model can be trained and validated using the training set and validation set to obtain the trained positioning error detection model (step 204). After obtaining the trained positioning error detection model, the field of view area can be determined using the vehicle's positioning data to determine the road mark information (step 205). Then, the first feature information is determined based on the combined navigation reference information corresponding to the road mark information (step 206). The first feature information is input into the trained positioning error detection model to obtain the positioning error detection result (step 207).

[0149] For example, such as Figure 4As shown, the training set and validation set can be determined based on the first road data generated by the vehicle and the high-precision ground truth map data. Then, the initial positioning error detection model can be trained using the training set and validation set. Thus, the positioning error detection model can be used to detect the positioning error of the vehicle and obtain the positioning error detection result.

[0150] For example, such as Figure 5 As shown, the vehicle may include a combined navigation module and a data vectorization module. Thus, the positioning error detection device can use the combined navigation reference information output by the combined navigation module and the road marking information output by the data vectorization module to determine the first feature information, and then input the first feature information into the trained LSTM model to obtain the positioning error detection result.

[0151] In some embodiments of this application, the above-described positioning error detection method can be used to evaluate the location accuracy of map fragments collected in a crowdsourced manner before map making. The location accuracy can be reflected by the positioning error, so that map making can be based on the positioning error detection results, thereby improving the accuracy of map making. The map fragments can be stored in the GeoJSON structured file format.

[0152] In some embodiments of this application, when determining the first feature information, the positioning error detection device can use a sliding time window with a step size of k to process the combined navigation reference information, road marking information, etc. in the time series to determine the first feature information.

[0153] In some embodiments of this application, signal features are mined from the original road data, and then the positioning error detection result is obtained through a deep learning positioning error detection model. This allows for the determination of whether the location accuracy of the crowdsourced map elements to be evaluated meets the standards, thereby achieving quality evaluation of the crowdsourced map elements and supporting quality control strategies in crowdsourced mapping technology routes.

[0154] In some embodiments of this application, a no-reference indirect evaluation method is adopted, which combines multi-source sensor data fusion, feature selection and deep learning algorithms to evaluate the location accuracy of map elements; long and short time series models are used as modeling tools, which can efficiently capture temporal features and complex relationships to improve the accuracy and robustness of data quality evaluation.

[0155] In some embodiments of this application, the first road data acquired by the vehicle may include at least road elements such as lane lines, arrows, and road boundaries; the positioning error detection device can remove low-quality or invalid data from the first road data, including filtering of road marking patterns and data attributes, thereby ensuring the accuracy and completeness of the obtained second road data and improving the training effect of the positioning error detection model.

[0156] In some embodiments of this application, the minimum bounding rectangle algorithm can be used to extract the bounding rectangle of each road marker to be processed in the first road data. The MBR algorithm is based on the rotating caliper algorithm, which finds the bounding rectangle with the smallest area by traversing the rotation angle of the polygon vertices.

[0157] In some embodiments of this application, the four vertices of the circumscribed rectangle can be determined first, and then the lengths of the four sides of the circumscribed rectangle can be determined based on the four vertices. Then, the lengths of the four sides are sorted according to their lengths, the longest side is defined as the long side, and the remaining second longest side is defined as the short side. Then, the side lengths and included angles can be determined based on the methods of the aforementioned formulas (9) to (13), and the road marking pattern is screened based on the side lengths and included angles, and data that does not conform to the road marking pattern is removed.

[0158] In some embodiments of this application, data that does not conform to the road marking type can also be removed based on the category attribute identification information of the first road data to ensure the integrity and consistency of the data input into the model.

[0159] In some embodiments of this application, after obtaining the second road data, the road surface features in the second road data, such as arrows and lane lines on the road surface, can be used to calculate the position error with the ground truth features in the ground truth map data corresponding to the road surface features. That is, by calculating the Euclidean distance between the geometric centroid of the road surface feature and the centroid of the ground truth feature, the shortest distance is selected as the position error, and these position errors are used as a dataset for model training and validation.

[0160] In some embodiments of this application, when performing positioning error detection, the positioning error detection device can obtain a prediction result of the positioning error at a future time k based on the vehicle's positioning data and environmental data over a certain period of time, as well as the trained positioning error detection model; that is, the positioning error detection device can predict the future positioning error based on historical data.

[0161] In summary, this application proposes a positioning error prediction scheme suitable for crowdsourced high-precision map data by applying the sequence dependency capture of time series models and multi-source data fusion technology. This scheme can effectively capture complex spatiotemporal feature relationships, significantly improve the accuracy and environmental adaptability of positioning quality assessment through deep learning models, and enhance the prediction accuracy of positioning errors for high-precision map data.

[0162] This application provides a positioning error detection method. The positioning error detection device can acquire vehicle positioning data and environmental data. The environmental data represents image data of the vehicle's surrounding environment during driving. Road marking information is determined based on the environmental data and positioning data. This road marking information represents information about road markings corresponding to the vehicle's driving trajectory. To-be-processed feature information is determined based on the road marking information, and the positioning error detection result is determined based on the to-be-processed feature information and a trained positioning error detection model. Therefore, the positioning error detection device can receive the vehicle's own positioning data and the image data of its surrounding environment in real time during vehicle operation. It can then use the positioning data and environmental data to determine the road marking information corresponding to the vehicle's driving trajectory, i.e., the information of road markings encountered during driving. Based on this road marking information, to-be-processed feature information can be extracted, and the positioning error detection result can be determined using the to-be-processed feature information and a trained positioning error detection model. This achieves effective detection of vehicle positioning errors, thereby improving the accuracy of vehicle positioning.

[0163] Based on the above embodiments, in another embodiment of this application... Figure 6 Schematic diagram of the composition structure of the positioning error detection device provided in the embodiments of this application Figure 1 ,like Figure 6 As shown, the positioning error detection device 1 provided in this application embodiment may include an acquisition unit 11, a determination unit 12, a positioning error detection unit 13, and a training unit 14.

[0164] The acquisition unit 11 can be used to acquire the vehicle's positioning data and environmental data; wherein, the environmental data represents image data of the vehicle's surrounding environment during the driving process.

[0165] The determining unit 12 can be used to determine road marking information based on environmental data and positioning data; wherein, the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0166] The positioning error detection unit 13 can be used to determine the feature information to be processed based on the road marking information, and to determine the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model.

[0167] In some embodiments of this application, the determining unit 12 can also be used to determine the field of view area of ​​the vehicle based on positioning data; and to determine road marking information in the field of view area based on the spatial location information of environmental data; wherein the field of view area represents the area of ​​the vehicle's field of view during driving.

[0168] In some embodiments of this application, the determining unit 12 can also be used to determine the driving direction of the vehicle based on the positioning data; and to determine the coordinate information of the field of view boundary point based on the position information of the vehicle's camera and the driving direction; and to determine the field of view region based on the position information of the vehicle's camera and the coordinate information of the field of view boundary point.

[0169] In some embodiments of this application, the positioning error detection unit 13 can also be used to obtain the combined navigation reference information corresponding to the road marking information, and determine the first feature information based on the combined navigation reference information; determine the second feature information based on the road marking information and the camera's position information; determine the third feature information based on the road marking information and the positioning data; and determine the feature information to be processed based on the first feature information, the second feature information and the third feature information; wherein, the combined navigation reference information includes at least one of the solution status information, satellite search number information and speed information output by the vehicle's combined navigation module.

[0170] The training unit 14 can be used to acquire first road data before the positioning error detection unit 13 determines the feature information to be processed based on the road marking information and determines the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model; and to perform data filtering processing on the first road data to obtain second road data; wherein the data filtering processing includes at least one of road marking pattern filtering processing and data attribute filtering processing; and to train the initial positioning error detection model based on the second road data and the ground truth map data to obtain the trained positioning error detection model.

[0171] Training unit 14 can also be used to determine the circumscribed rectangle of the road markers to be processed in the first road data; and to perform road marker morphology filtering based on the first morphological information of the circumscribed rectangle and the second morphological information corresponding to the road markers to be processed in the ground truth map data to obtain the second road data.

[0172] Training unit 14 can also be used to determine the unprocessed identification information in the first road data that does not meet the preset identification conditions based on the category attribute identification information of the first road data; and to remove the road data corresponding to the unprocessed identification information in the first road data to complete the data attribute filtering process and obtain the second road data.

[0173] In the embodiments of this application, further, Figure 7 Schematic diagram of the composition structure of the positioning error detection device provided in the embodiments of this application Figure 2 ,like Figure 7 As shown, the positioning error detection device 1 provided in this application embodiment may further include a processor 15 and a memory 16 storing executable instructions of the processor 15; further, the positioning error detection device 1 may further include a communication interface 17 and a bus 18 for connecting the processor 15, the memory 16 and the communication interface 17.

[0174] In the embodiments of this application, the processor 15 can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, and microprocessor. It is understood that for different devices, the electronic device used to implement the above-mentioned processor function can also be other types, and this application embodiment does not specifically limit this. The positioning error detection device 1 may further include a memory 16, which can be connected to the processor 15. The memory 16 is used to store executable program code, which includes computer operation instructions. The memory 16 may include high-speed RAM memory and may also include non-volatile memory, such as at least two disk drives.

[0175] In embodiments of this application, bus 18 is used to connect communication interface 17, processor 15, and memory 16, as well as the mutual communication between these devices.

[0176] In embodiments of this application, memory 16 is used to store instructions and data.

[0177] Furthermore, in the embodiments of this application, the processor 15 is used to acquire vehicle positioning data and environmental data; wherein, the environmental data represents image data of the vehicle's surrounding environment during driving.

[0178] Road marking information is determined based on environmental and location data; the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0179] The feature information to be processed is determined based on the road marking information, and the positioning error detection result of the vehicle is determined based on the feature information to be processed and the trained positioning error detection model.

[0180] In practical applications, the aforementioned memory 16 can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or a combination of the above types of memory, and provide instructions and data to the processor 15.

[0181] Furthermore, in this embodiment, the functional modules 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. The integrated unit can be implemented in hardware or as a software functional module.

[0182] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part 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.) or processor to execute all or part of the steps of the method of this embodiment. 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.

[0183] This application provides a positioning error detection device that can acquire vehicle positioning data and environmental data. The environmental data represents image data of the vehicle's surrounding environment during driving. Road marking information is determined based on the environmental data and positioning data. This road marking information represents information about road markings corresponding to the vehicle's driving trajectory. To-be-processed feature information is determined based on the road marking information, and the positioning error detection result is determined based on the to-be-processed feature information and a trained positioning error detection model. Therefore, the positioning error detection device can receive the vehicle's own positioning data and the image data of its surrounding environment in real time during vehicle operation. It can then use the positioning data and environmental data to determine the road marking information corresponding to the vehicle's driving trajectory, i.e., the information about road markings encountered during driving. Based on this road marking information, to-be-processed feature information can be extracted, and the positioning error detection result can be determined using the to-be-processed feature information and a trained positioning error detection model. This achieves effective detection of vehicle positioning errors, thereby improving the accuracy of vehicle positioning.

[0184] Specifically, the program instructions corresponding to a positioning error detection method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to a positioning error detection method in the storage media are read or executed by a positioning error detection device, the following steps are included:

[0185] Acquire vehicle location data and environmental data; among which, environmental data represents image data of the vehicle's surrounding environment during driving.

[0186] Road marking information is determined based on environmental and location data; the road marking information represents the information of road markings corresponding to the vehicle's driving trajectory.

[0187] The feature information to be processed is determined based on the road marking information, and the positioning error detection result of the vehicle is determined based on the feature information to be processed and the trained positioning error detection model.

[0188] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0189] This application is described with reference to schematic and / or block diagrams of implementations of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the schematic and / or block diagrams can be implemented by computer program instructions, and combinations of blocks in the schematic and / or block diagrams can be implemented. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the schematic and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0190] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in the implementation flow diagram. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0191] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0192] The above embodiments are merely preferred embodiments provided to fully illustrate this application, and the scope of protection of this application is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on this application are all within the scope of protection of this application.

Claims

1. A positioning error detection method, characterized in that, The method includes: Acquire vehicle positioning data and environmental data; wherein the environmental data represents image data of the vehicle's surrounding environment during driving. Road marking information is determined based on the environmental data and the positioning data; wherein the road marking information represents information about road markings corresponding to the vehicle's driving trajectory; The feature information to be processed is determined based on the road marking information, and the positioning error detection result of the vehicle is determined based on the feature information to be processed and the trained positioning error detection model.

2. The method according to claim 1, characterized in that, The process of determining road marking information based on the environmental data and the location data includes: The field of view region of the vehicle is determined based on the positioning data; wherein the field of view region represents the area of ​​the field of view of the vehicle during driving. The road marking information in the field of view area is determined based on the spatial location information of the environmental data.

3. The method according to claim 2, characterized in that, Determining the field of view region of the vehicle based on the positioning data includes: The vehicle's driving direction is determined based on the positioning data; The coordinate information of the field of view boundary point is determined based on the position information of the vehicle's camera and the driving direction; The field of view region is determined based on the location information of the vehicle's camera and the coordinate information of the field of view boundary point.

4. The method according to claim 3, characterized in that, The step of determining the feature information to be processed based on the road marking information includes: Obtain the combined navigation reference information corresponding to the road marking information, and determine the first feature information based on the combined navigation reference information; wherein, the combined navigation reference information includes at least one of the solution status information, satellite search number information, and speed information output by the vehicle's combined navigation module; The second feature information is determined based on the road marking information and the camera's location information; The third feature information is determined based on the road marking information and the positioning data; The feature information to be processed is determined based on the first feature information, the second feature information, and the third feature information.

5. The method according to any one of claims 1 to 4, characterized in that, Before determining the feature information to be processed based on the road marking information, and determining the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model, the method further includes: Obtain the first road data; The first road data is subjected to data filtering processing to obtain the second road data; wherein, the data filtering processing includes at least one of road mark pattern filtering processing and data attribute filtering processing; The initial positioning error detection model is trained based on the second road data and the ground truth map data to obtain the trained positioning error detection model.

6. The method according to claim 5, characterized in that, The step of filtering the first road data to obtain the second road data includes: Determine the bounding rectangle of the road marker to be processed in the first road data; Based on the first shape information of the circumscribed rectangle and the second shape information corresponding to the road mark to be processed in the ground truth map data, the road mark shape is filtered to obtain the second road data.

7. The method according to claim 5, characterized in that, The step of filtering the first road data to obtain the second road data includes: Based on the category attribute identification information of the first road data, identify the identification information to be processed in the first road data that does not meet the preset identification conditions; The road data corresponding to the identification information to be processed in the first road data is removed to complete the data attribute filtering process and obtain the second road data.

8. A positioning error detection device, characterized in that, It includes an acquisition unit, a determination unit, and a positioning error detection unit; The acquisition unit is used to acquire vehicle positioning data and environmental data; wherein the environmental data represents image data of the vehicle's surrounding environment during driving. The determining unit is used to determine road marking information based on the environmental data and the positioning data; wherein the road marking information represents information about road markings corresponding to the vehicle's driving trajectory; The positioning error detection unit is used to determine the feature information to be processed based on the road marking information, and to determine the positioning error detection result of the vehicle based on the feature information to be processed and the trained positioning error detection model.

9. A positioning error detection device, characterized in that, The positioning error detection device includes a processor and a memory storing executable instructions of the processor; when the instructions are executed by the processor, the method described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images