Road edge tracking method and device and storage medium

Abstract

The invention discloses a road edge tracking method and device and a storage medium, and the method comprises the steps: carrying out the rasterization of a plurality of detection road edge points of a road at a current moment, and obtaining a corresponding raster map; clustering by using the positions of the grids where the detection road edge points are located in the grid map to obtain a plurality of target road edge point sets; the plurality of target road edge point sets are used for determining a plurality of first road edge lines at the current moment; and obtaining a second fusion road line at the current moment based on the first fusion road line at the historical moment and the plurality of first road lines at the current moment. In this way, the road edge tracking efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent transportation technology, and in particular to a curb tracking method, device and storage medium. Background Art

[0002] The perception of the surrounding environment during vehicle driving is the basis for realizing intelligent assisted driving and unmanned driving. Curb detection technology is an important link in realizing intelligent path planning and decision-making control of vehicles, and is also the basis for realizing assisted driving such as lane keeping assist and lane departure warning.

[0003] The core of curb tracking is to accurately and in real time detect and track the curb of the road, so as to provide vehicles with real-time driving boundary information and ensure that the vehicle drives within the specified road area. Therefore, how to improve the efficiency of curb detection is crucial for real-time curb tracking. Summary of the invention

[0004] The main technical problem solved by the present application is to provide a roadside tracking method, device and storage medium, which can improve the detection efficiency of the roadside line and further improve the tracking efficiency of the roadside.

[0005] In order to solve the above technical problems, a technical solution adopted in the present application is: to provide a roadside tracking method, the method comprising: rasterizing multiple detected roadside points of the road at the current moment to obtain a corresponding grid map; clustering the positions of the grids where each detected roadside point is located in the grid map to obtain multiple target roadside point sets; the multiple target roadside point sets are used to determine several first roadside lines at the current moment; based on the first fused roadside line at the historical moment and several first roadside lines at the current moment, a second fused roadside line at the current moment is obtained.

[0006] Among them, clustering is performed using the positions of the grids where each detected roadside point is located in the grid map to obtain multiple target roadside point sets, including: selecting an unsearched roadside point from the multiple detected roadside points as the first roadside point; searching for detected roadside points of the same category as the first roadside point from the first grid search range to obtain an initial roadside point set corresponding to the first roadside point; the first grid search range is determined based on the first position of the first grid where the first roadside point is located in the grid map, and is used to represent the grid range to be searched in the grid map; based on the initial roadside point set corresponding to the first roadside point, a target roadside point set corresponding to the first roadside point is determined; the target roadside point set includes all detected roadside points of the same category as the first roadside point among all detected roadside points; repeating the above steps until all detected roadside points in the grid map have been searched, and obtaining target roadside point sets corresponding to each first roadside point.

[0007] Wherein, based on the initial roadside point set corresponding to the first roadside point, determining the target roadside point set corresponding to the first roadside point includes: taking each detected roadside point except the first roadside point in the initial roadside point set as a second roadside point; for each second roadside point, searching for a detected roadside point of the same category as the second roadside point from a second grid search range, and adding the searched detected roadside point to the initial roadside point set; the second grid search range is determined based on a second position of the second grid where the second roadside point is located in the grid map; and dividing each searched detected roadside point into The method further comprises the following steps: respectively selecting each second roadside point as a new second roadside point, and returning to the step of searching for detection roadside points of the same category as the second roadside point in the second grid search range for each second roadside point, and adding the searched detection roadside points to the initial roadside point set, until there are no detection roadside points of the same category as the second roadside points in the second grid search range, or the number of detection roadside points of the same category as the second roadside points searched is less than the set threshold, and taking the last determined initial roadside point set as the target roadside point set corresponding to the first roadside point.

[0008] Among them, the grid map includes a plurality of grids distributed in rows and columns, and the position of the grid where the detection roadside point is located in the grid map includes the row and column values ​​of the grid in the grid map; the detection roadside points of the same category as the first roadside point among all the detection roadside points include: the second roadside points located in the first grid search range corresponding to the first roadside point, and the detection roadside points located in the second grid search range corresponding to the second roadside points; the first grid search range is determined by using the first row and column values ​​of the first grid where the first roadside point is located in the grid map and the preset search range value; the second grid search range is determined based on the second row and column values ​​of the second grid where the second roadside point is located in the grid map and the preset search range value.

[0009] Among them, the preset search range value is smaller than the road width; the step of determining the grid search range includes: determining the number of grid searches using the preset search range value and the resolution of the grid map; determining the grid search range based on the target row and column values ​​of the target grid in the grid map and the number of grid searches; wherein the grid search range is any one of a first grid search range and a second grid search range, in response to the grid search range being the first grid search range, the target grid represents the first grid where the first roadside point is located; in response to the grid search range being the second grid search range, the target grid represents the second grid where the second roadside point is located.

[0010] The target row and column values ​​include the target row value and the target column value of the target grid in the grid map; based on the target row and column values ​​of the target grid in the grid map and the grid search quantity, determining the grid search range includes: in response to a first difference between the target row value and the grid search quantity being not less than a minimum row value, taking a first numerical value corresponding to the first difference as the minimum grid search row value, otherwise, taking the minimum row value as the minimum grid search row value; in response to a first summation result of the target row value and the grid search quantity being not greater than a maximum row value, taking a second numerical value corresponding to the first summation result as the maximum grid search row value, otherwise, taking the maximum row value as the maximum grid search row value; in response to a first summation result of the target row value and the grid search quantity being not greater than a maximum row value, taking a second numerical value corresponding to the first summation result as the maximum grid search row value, otherwise, taking the maximum row value as the maximum grid search row value; The second difference between the search quantities is not less than the minimum column value, and the third value corresponding to the second difference is used as the minimum grid search column value, otherwise, the minimum column value is used as the minimum grid search column value; in response to the second summation result of the target column value and the grid search quantity being not greater than the maximum column value, the fourth value corresponding to the second summation result is used as the maximum grid search column value, otherwise, the maximum column value is used as the maximum grid search column value; the minimum grid search row value, the maximum grid search row value, the minimum grid search column value and the maximum grid search column value are respectively used as search boundary values ​​to construct a grid search range; wherein the minimum row value, the maximum row value, the minimum column value and the maximum column value are respectively the map boundary values ​​of the grid map.

[0011] Among them, determining a number of first roadside lines at the current moment includes: obtaining second roadside lines determined based on each target roadside point set; determining each target roadside line pair that meets the splicing condition from each second roadside line; splicing based on two second roadside lines in each target roadside line pair to obtain a corresponding spliced ​​roadside line; taking each spliced ​​roadside line as a first roadside line, and taking each remaining second roadside line that does not meet the splicing condition in each second roadside line as a first roadside line, to obtain a number of first roadside lines.

[0012] Among them, determining each target roadside line pair that meets the splicing condition from each second roadside line includes: selecting each candidate roadside line pair to be matched from each second roadside line; for each candidate roadside line pair, determining whether the two second roadside lines meet the first splicing condition based on the first position difference between the two second roadside lines in the candidate roadside line pair; in response to satisfying the first splicing condition, determining that the two second roadside lines meet the splicing condition; or, in response to satisfying the first splicing condition, determining whether the two second roadside lines meet the second splicing condition based on the detection result of the obstacle at the current moment; in response to the two second roadside lines satisfying the second splicing condition, determining that the two second roadside lines meet the splicing condition; and using the two second roadside lines that meet the splicing condition as the two second roadside lines in a target roadside line pair.

[0013] The first position difference includes at least one of a lateral position difference and an angle difference; the first splicing condition includes that the lateral position difference is smaller than a lateral threshold, and / or the angle difference is smaller than an angle threshold.

[0014] Among them, determining the lateral difference includes: taking one of the two second roadside lines as the first target roadside line, taking the other as the second target roadside line, and selecting a target point from the first target roadside line; obtaining an associated target point of the target point, the associated target point is obtained by extending the second target roadside line, and the longitudinal position difference between the target point and the associated target point is less than a preset difference threshold; taking the lateral position difference between the target point and the associated target point as the lateral position difference between the two second roadside lines; determining the angle difference includes: obtaining the slope representation value of each of the two second roadside lines; taking the difference between the slope representation values ​​corresponding to the two second roadside lines as the angle difference between the two second roadside lines.

[0015] Among them, the slope characterization value corresponding to the first target roadside line is the first slope value of the curve equation corresponding to the first target roadside line at the target point, and the slope characterization value corresponding to the second target roadside line is the second slope value of the curve equation corresponding to the second target roadside line at the associated target point; the curve equation corresponding to each second roadside line is obtained by curve fitting the corresponding target roadside point set; and / or, the target point is a detected roadside point on the first target roadside line close to the second target roadside line.

[0016] Among them, the detection result of the obstacle includes one of a first detection result in which there is no obstacle and a second detection result in which there is an obstacle; based on the detection result of the obstacle at the current moment, determining whether the two second roadside lines meet the second splicing condition, including: in response to the detection result being the first detection result, determining that the two second roadside lines do not meet the second splicing condition; in response to the detection result being the second detection result, determining the second position difference between the occlusion area corresponding to the obstacle and the two second roadside lines; based on the second position difference, determining whether the two second roadside lines meet the second splicing condition.

[0017] Among them, the second detection result includes the first position of the obstacle detection frame, and the occlusion area is determined using the first position of the obstacle detection frame; determining the second position difference between the occlusion area corresponding to the obstacle and the two second roadside lines, including: determining the first intersection points of the two second roadside lines or the corresponding extension lines with the occlusion area respectively; taking the longitudinal position difference between the target roadside point on each second roadside line and the first intersection point of the corresponding second roadside line and the occlusion area as the second position difference; the second splicing condition includes: each longitudinal position difference is not greater than the longitudinal difference threshold.

[0018] Among them, the second roadside line includes a starting roadside point and an ending roadside point, and the starting roadside point and the ending roadside point are determined based on the position sorting of each detected roadside point in the target roadside point set corresponding to the second roadside line; the target roadside point on the front roadside line of the two second roadside lines is the starting roadside point on the front roadside line, and the target roadside point on the rear roadside line is the ending roadside point on the rear roadside line.

[0019] Among them, the occlusion area corresponding to the obstacle is determined by using the second positions corresponding to the four corner points of the obstacle detection frame, and the second positions corresponding to the corner points are determined based on the first position of the obstacle detection frame; determining the occlusion area includes: in response to the obstacle being located on the same side of the vehicle's forward direction, based on the preset position point on the vehicle at the current moment and the second position of each corner point, constructing connecting lines between the preset position point and each corner point; obtaining the angle between each connecting line and the vehicle's forward direction, and taking the connecting line corresponding to the smallest of each angle as the first connecting line, and taking the connecting line corresponding to the largest of each angle as the second connecting line; taking the area between the first connecting line and the second connecting line as the occlusion area; the first intersection points of the two second roadside lines or the corresponding extension lines with the occlusion area include: the second intersection points of the two second roadside lines or the corresponding extension lines with the first connecting line, and the third intersection points of the two second roadside lines or the corresponding extension lines with the second connecting lines.

[0020] Among them, obtaining the second roadside lines determined based on each target roadside point set respectively includes: performing a first curve fitting using each target roadside point set respectively to obtain the second roadside lines corresponding to each target roadside point set; and / or, splicing two second roadside lines in each target roadside line pair to obtain a corresponding spliced ​​roadside line, including: for each target roadside line pair, merging the target roadside point sets corresponding to each second roadside line in the target roadside line pair to obtain a merged roadside point set; performing a second curve fitting on the merged roadside point set to obtain a spliced ​​roadside line.

[0021] Among them, the number of first fused road lines is at least one; based on the first fused road lines at the historical moment and several first road lines at the current moment, the second fused road line at the current moment is obtained, including: based on the driving data of the vehicle at the historical moment, predicting that each first fused road line at the historical moment corresponds to the third fused road line at the current moment; associating each third fused road line with several first road lines to determine each road line association pair that is successfully associated; for each road line association pair, the third fused road line and the first road line in the road line association pair are fused to obtain the second fused road line at the current moment.

[0022] Among them, each third fused roadside line is associated with several first roadside lines to determine each successfully associated roadside line association pair, including: obtaining the lateral error between each third fused roadside line and each first roadside line respectively; and associating the roadside lines based on each lateral error to obtain each successfully associated roadside line association pair.

[0023] Among them, roadside lines are associated based on each lateral error to obtain each successfully associated roadside line association pair, including: based on the lateral error between each third fused roadside line and each first roadside line, a cost matrix for characterizing each lateral error is constructed; roadside lines are paired based on the cost matrix, and a roadside line target matching combination is selected from a number of roadside line candidate matching combinations; each roadside line candidate matching combination is a pairing combination including each third fused roadside line and each first roadside line; from each roadside line matching pair in the roadside line target matching combination, each roadside line association pair that meets the preset requirements is selected; the lateral error between the third fused roadside line and the corresponding first roadside line in each roadside line association pair is less than a preset threshold.

[0024] Among them, the sum of the first errors of all roadside matching pairs in the roadside target matching combination is less than the sum of the second errors of all roadside matching pairs in each roadside candidate matching combination; the sum of the errors is the sum of the lateral errors corresponding to all roadside matching pairs; after the roadsides are associated based on the lateral errors to obtain the successfully associated roadside associated pairs, the method also includes: determining whether there is a third fused roadside that has not been successfully associated among each third fused roadside; in response to the existence of the third fused roadside that has not been successfully associated, the third fused roadside that has not been successfully associated is used as the fourth fused roadside, and in response to If there is no first roadside line successfully associated with the fourth fused roadside line in the first preset number of associations in the future, the fourth fused roadside line is deleted; and / or, and / or, it is determined whether there is a first roadside line that has not been successfully associated among the first roadside lines; in response to the existence of a first roadside line that has not been successfully associated, the unsuccessfully associated first roadside line is used as a candidate tracking roadside line; in response to the existence of a first roadside line associated with the candidate tracking roadside line in the second preset number of associations in the future, the candidate tracking roadside line is determined to be a tracking roadside line; wherein the first fused roadside line and the third fused roadside line both belong to tracking roadside lines.

[0025] Among them, searching for detection roadside points of the same category as the target detection roadside points from the grid search range includes: determining the number of detection roadside points in the grid search range; in response to the number being greater than a preset number, treating each detection roadside point in the grid search range as a detection roadside point of the same category as the target detection roadside point; in response to the number being less than or equal to the preset number, treating each detection roadside point in the grid search range as a noise point; wherein the grid search range is any one of a first grid search range and a second grid search range, in response to the grid search range being the first grid search range, the target detection roadside point represents the first roadside point; in response to the grid search range being the second grid search range, the target detection roadside point represents the second roadside point.

[0026] To solve the above technical problems, another technical solution adopted in the present application is: to provide an electronic device, comprising a memory and a processor coupled to each other, the memory storing program instructions; the processor is used to execute the program instructions stored in the memory to implement the above method.

[0027] In order to solve the above technical problem, another technical solution adopted by the present application is: providing a computer-readable storage medium for storing program instructions, which can be executed to implement the above method.

[0028] The above scheme performs rasterization processing on multiple detected curb points of the road at the current moment to obtain a corresponding grid map. The detected curb points can be placed in the grid map, and then clustered using the position of the grid where each detected curb point is located in the grid map to obtain multiple target curb point sets for determining several roadside lines at the current moment. Since the present application can directly use the position of the grid for clustering without considering the position of each detected curb point in the grid, compared to the method that needs to use the position of each point for clustering, the method of clustering using the grid position in the present application is conducive to improving the efficiency of clustering to obtain multiple target curb point sets, that is, it can improve the efficiency of determining the roadside line, and thus is conducive to improving the subsequent curb tracking efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flowchart of an embodiment of a roadside tracking method provided by the present application;

[0030] Figure 2 yes Figure 1 The flowchart of step S12 is shown as an embodiment;

[0031] Figure 3 yes Figure 2 The flowchart of step S23 is shown as an embodiment;

[0032] Figure 4 It is a flowchart of an embodiment of determining the association pairs along each road provided by the present application;

[0033] Figure 5 It is a flowchart of an embodiment of determining several first routes at the current moment provided by the present application;

[0034] Figure 6 is a schematic diagram of an embodiment of determining a target roadside line according to the present application;

[0035] Figure 7 is a schematic diagram of a flow chart of an embodiment of determining a second position difference provided by the present application;

[0036] Figure 8 It is a flowchart of an embodiment of determining an occlusion area of ​​an obstacle provided by the present application;

[0037] Fig. 9 It is a schematic diagram of a framework of an embodiment of an electronic device provided by the present application;

[0038] Fig.10 It is a schematic diagram of the framework of the computer-readable storage medium provided by the present application;

[0039] Fig.11 It is a schematic diagram of the framework of an embodiment of a curb tracking device provided in the present application. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail below with reference to the accompanying drawings and examples.

[0041] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0042] See also Figure 1 , Figure 1 1 is a flow chart of an embodiment of the roadside tracking method provided by the present application. It should be noted that if there are substantially the same results, this embodiment does not necessarily refer to the embodiment of the present invention. Figure 1 The process sequence shown is limited. Figure 1 As shown, this embodiment includes:

[0043] S11: rasterizing a plurality of detected roadside points of the road at the current moment to obtain a corresponding raster map.

[0044] This embodiment is used to perform clustering using the positions of grids in a grid map to obtain target roadside point sets, so as to improve the clustering efficiency of detecting roadside points.

[0045] In this embodiment, the plurality of detected roadside points are obtained by detecting the collected road image. For example, the road image is obtained by semantic segmentation, or by detecting the road image using a relevant detection algorithm. The road image is, for example, a point cloud image collected by a laser radar, or an image of the roadside collected in real time by an image acquisition device.

[0046] Rasterization of multiple detection roadside points is performed, for example: first determine the range and resolution of the grid map to divide the grid into rows and columns, and then assign each detection roadside point to a corresponding grid according to the position of each detection roadside point to obtain a raster map after rasterization.

[0047] The range and resolution of the grid map constructed specifically can be set according to actual needs. For example, a 200×120 grid map with a resolution of 0.5m×0.5m is constructed 100 meters in front of the vehicle and 30 meters on the left and right.

[0048] According to the position of each detection roadside point, each detection roadside point is assigned to a corresponding grid, including: according to the ratio of the position of each detection roadside point to the resolution (take an integer), the position of the grid where each detection roadside point is located in the grid map, that is, the row and column values ​​of the grid in the grid map. If the row value of the grid calculated using the position of a certain detection roadside point is greater than the maximum row value of the grid map, or the calculated column value of the grid is greater than the maximum column value in the grid map, it means that the position of the detection roadside point is not within the range of the pre-established grid map, and the detection roadside point is not assigned to the grid map.

[0049] S12: clustering is performed using the positions of the grids where the detected roadside points are located in the grid map to obtain a plurality of target roadside point sets; the plurality of target roadside point sets are used to determine a plurality of first roadside lines at the current moment.

[0050] In this embodiment, the purpose of clustering is to cluster the detected roadside points with a short distance into a target roadside point set. Correspondingly, the roadside points in a target roadside point set can be regarded as roadside points belonging to the same category. The categories described in this application are only used to refer to the distance between the detection points, and are not used to distinguish different roadside lines on the road. Among them, it can be understood that if there is no intersection on the current road, and there is no obstacle blocking the roadside line, a plurality of detected roadside points will generally be divided into two target roadside point sets, and each target roadside point set represents a roadside line; but if there is an intersection on the current road, or there is an obstacle blocking the roadside line, a plurality of detected roadside points will be divided into at least three target roadside point sets, among which there may be two target roadside point sets belonging to the roadside point sets on the same roadside (line) on the road.

[0051] For example, a current viewing angle of the vehicle is blocked by an obstacle, and a corresponding complete roadside line is divided into two sections of the roadside line due to the blockage of the obstacle, and each section of the roadside line corresponds to a target roadside point set.

[0052] Therefore, in this embodiment, the number of target roadside point sets is generally greater than or equal to the number of first roadside lines, wherein each first roadside line is a complete roadside line. For example, when a complete roadside line is divided into two sections of roadside lines by an obstacle, the corresponding roadside line after the two discontinuous roadside lines are supplemented and connected is a first roadside line.

[0053] In this embodiment, a roadside point that has not been clustered searched can be randomly selected in the constructed grid map as the first roadside point (which can be regarded as a basic roadside point for searching roadside points belonging to the same category), and then the detection roadside points of the same category as the first roadside point are searched from the first grid search range corresponding to the first roadside point to obtain the initial roadside point set corresponding to the first roadside point, and then the target roadside point set corresponding to the first roadside point is determined based on the initial roadside point set corresponding to the first roadside point. Then, another first roadside point is selected from the remaining detection roadside points that have not been clustered searched in the current grid map, and the same method is used to perform cluster search to obtain the target roadside point set corresponding to the first roadside point, until all the detection roadside points in the grid map have been clustered searched, and the target roadside point set corresponding to each first roadside point is obtained.

[0054] In some embodiments, in order to facilitate the determination of which detection roadside points in the grid map have been searched and which detection roadside points have not been searched, a first mark representing the searched detection roadside points and a second mark representing the unsearched detection roadside points may be pre-set. Specifically, before the first search, all detection roadside points may be marked with the second mark (e.g., number 0), and after the corresponding detection roadside points have been searched, the corresponding search mark may be adjusted to the first mark (e.g., 1). The first mark and the second mark may be set according to actual needs, and may be, but not limited to, numbers, letters, or punctuation marks.

[0055] Specifically, see Figure 2 , Figure 2 yes Figure 1 The flowchart of an embodiment of step S12 is shown. In this embodiment, step S12 further includes:

[0056] S21: Selecting an unsearched roadside point from a plurality of detected roadside points as the first roadside point.

[0057] S22: Searching for detected roadside points of the same category as the first roadside point in the first grid search range to obtain an initial roadside point set corresponding to the first roadside point.

[0058] In one embodiment, the number of detected roadside points in the first grid search range may be ignored, and each detected roadside point in the first grid search range is regarded as a detected roadside point of the same category as the first roadside point.

[0059] In another embodiment, in order to reduce the influence of noise points on the subsequent determination of the first roadside line, the detected roadside points in the first grid search range that are of the same category as the first roadside point can be determined based on the number of detected roadside points in the first grid search range. Specifically, if the number of detected roadside points in the first grid search range is less than or equal to a preset number, each detected roadside point in the first grid search range is regarded as a noise point, and is not determined as a detected roadside point of the same category as the first roadside point; on the contrary, if the number of detected roadside points in the first grid search range is greater than a preset number, each detected roadside point in the first grid search range is regarded as a detected roadside point of the same category as the first roadside point, and an initial roadside point set corresponding to the first roadside point is obtained.

[0060] It should be noted that, in this embodiment, if step S22 fails to search for a detected roadside point of the same category as the first roadside point in the first grid search range, the process returns to step S21 and re-executes step S21 and subsequent steps.

[0061] S23: Based on the initial roadside point set corresponding to the first roadside point, determine the target roadside point set corresponding to the first roadside point; the target roadside point set includes all detected roadside points that are of the same category as the first roadside point among all detected roadside points.

[0062] S24: Repeat the above steps until all detected roadside points in the grid map have been searched, and a target roadside point set corresponding to each first roadside point is obtained.

[0063] The first grid search range corresponding to each first roadside point is determined by using the first position of the first grid where the first roadside point is located in the grid map, and is used to represent the grid range to be searched in the grid map. Each target roadside point set includes all detected roadside points that are of the same category as the corresponding first roadside point.

[0064] In summary, when clustering is performed using a grid map in this embodiment, each target roadside point set is obtained in sequence. For example, a first roadside point A is first selected from the currently unsearched detection roadside points, and the first roadside point A is used for clustering to obtain a target roadside point set of the same category as the first roadside point A. Then, a first roadside point B is selected from the currently searched detection roadside points, and the first roadside point B is used for clustering to obtain a target roadside point set of the same category as the first roadside point B. Clustering is performed in the same manner until there are no unsearched detection roadside points among all the detection roadside points.

[0065] In this embodiment, step S23 determines the target roadside point set corresponding to the first roadside point based on the initial roadside point set corresponding to the first roadside point, including at least the following two methods:

[0066] Method 1: directly determine the initial roadside point set corresponding to the first roadside point as the target roadside point set corresponding to the first roadside point.

[0067] Method 2: Based on the initial roadside point set corresponding to the first roadside point, the search area is expanded using the newly added detected roadside points in the current set to expand the number of detected roadside points in the initial roadside point set until there are no detected roadside points in the expanded area.

[0068] Specifically, see Figure 3 , Figure 3 yes Figure 2 The flowchart of step S23 of an embodiment is shown. In this embodiment, based on the initial roadside point set corresponding to the first roadside point, determining the target roadside point set corresponding to the first roadside point includes:

[0069] S31: taking each detected roadside point in the initial roadside point set except the first roadside point as a second roadside point.

[0070] S32: For each second roadside point, search for a detection roadside point of the same category as the second roadside point from the second grid search range, and add the searched detection roadside point to the initial roadside point set; the second grid search range is determined based on the second position of the second grid where the second roadside point is located in the grid map.

[0071] In one embodiment, each detected roadside point except the first roadside point in the initial roadside point set may be directly used as a second roadside point to expand the search area.

[0072] In another embodiment, in order to avoid duplication of the search area, each detected roadside point in other grids except the grid where the first roadside point is located can be used as the second roadside point. That is, only the initial roadside point set is used, and each detected roadside point in other grids except the grid where the first roadside point is located is used as the second roadside point to expand the search area.

[0073] After determining the extended area (second grid search range), search for detected roadside points of the same category as the second roadside point from the second grid search range, and add the searched detected roadside points to the initial roadside point set to expand the number of roadside points in the initial roadside point set.

[0074] In this embodiment, the number of detected roadside points in the second grid search range may be ignored, and each detected roadside point in the second grid search range may be regarded as a detected roadside point of the same category as the second roadside point. The detected roadside points of the same category as the corresponding second roadside point in the second grid search range may also be determined based on the number of detected roadside points in the second grid search range. Specifically, if the number of detected roadside points in the second grid search range is less than or equal to a preset number, each detected roadside point in the second grid search range is regarded as a noise point, and is not determined as a detected roadside point of the same category as the second roadside point; on the contrary, if the number of detected roadside points in the second grid search range is greater than a preset number, each detected roadside point in the second grid search range is regarded as a detected roadside point of the same category as the corresponding second roadside point, and the detected roadside points of the same category as the second roadside point are added to the initial roadside point set.

[0075] S33: Take each searched detection roadside point as a new second roadside point, and return to the step of searching for detection roadside points of the same category as the second roadside point from the second grid search range for each second roadside point, and adding the searched detection roadside points to the initial roadside point set, until there are no detection roadside points of the same category as the second roadside points in each second grid search range, or the number of detection roadside points of the same category as the second roadside points searched is less than a set threshold, and the last determined initial roadside point set is used as the target roadside point set corresponding to the first roadside point.

[0076] After adding the searched detection roadside points to the initial roadside point set in step S32, each newly searched detection roadside point in step S32 is used as a new second roadside point, and a new round of search area expansion is performed to further increase the number of detection roadside points in the initial roadside point set. Specifically, the process of the new round of search is the same as step S32 and step S33. Among them, in each round of search process, as long as a detection roadside point of the same category as the second roadside point is searched in the second grid search range corresponding to any second roadside point, the searched detection roadside point is added to the initial roadside point set, and a subsequent new round of search process is performed until there is no detection roadside point of the same category as each second roadside point in the second grid search range corresponding to each second roadside point, or the number of detection roadside points of the same category as each second roadside point is less than the set threshold, and the initial roadside point set determined last time is used as the target roadside point set corresponding to the first roadside point. The set threshold and the above-mentioned preset number may be the same or different. The specific set threshold and the above-mentioned preset number may be determined based on experience and are not specifically limited here.

[0077] It should be noted that the above-mentioned grid map includes grids distributed in a number of rows and columns, and the position of the grid where the roadside point is located in the grid map includes the row and column values ​​of the grid in the grid map. That is, the position of the grid in the grid map is represented by the row and column values, for example, the row and column of grid A in the grid map.

[0078] The categories of the detected roadside points in the target roadside point set corresponding to the first roadside point are the same. Figure 2 and Figure 3 It can be seen from the steps that all the detected roadside points that are of the same category as the first roadside point include: each second roadside point located in the first grid search range corresponding to the first roadside point, and each detected roadside point located in the second grid search range corresponding to each second roadside point.

[0079] Among them, the first grid search range is determined by using the first row and column values ​​of the first grid where the first roadside point is located in the grid map and the preset search range value; the second grid search range is determined based on the second row and column values ​​of the second grid where the second roadside point is located in the grid map and the preset search range value.

[0080] It should be noted that in order to use the above method to distinguish different roadside lines on the road, the preset search range value can be set to be smaller than the road width. The specific preset search range value is determined based on experience and is set based on the ability to cluster roadside points on different sections and sides of the roadside lines into different sets.

[0081] In one embodiment, the first grid search range corresponding to the first wayside point and the second grid search range corresponding to the second wayside point are determined in the same manner.

[0082] For the convenience of description, the first grid search range and the second grid search range are collectively referred to as the grid search range, and the grids where the first roadside point and the second roadside point are located are collectively referred to as the target grid. The steps of determining the grid search range include:

[0083] First, the number of grid searches is determined using the preset search range value and the resolution of the grid map.

[0084] Second, the grid search range is determined based on the target row and column values ​​of the target grid in the grid map and the number of grid searches.

[0085] The target row and column values ​​include the target row and column values ​​of the target grid in the grid map. The number of grid searches is the ratio of the preset search range value to the resolution of the grid map. The purpose of this embodiment is to use the grids within the preset search range value of the target grid as grids within the grid search range, and then perform subsequent searches for roadside points of the same category.

[0086] Exemplarily, the preset search range value (search radius) is 2.5 meters, and the purpose of the search is to use the first roadside point or the second roadside point as the search center, and all grids within a radius of 2.5 meters are grids within the grid search range. To facilitate the determination of the grid search range, the preset search range value and the resolution of the grid map can be used to determine the number of grid searches (here, the number of grid searches is used to represent the above search radius), and then the grid search range is determined according to the target row and column values ​​of the target grid in the grid map and the number of grid searches.

[0087] Specifically, in order to ensure that all grids within the grid search range are located in the grid map, the map boundary value of the grid map can be determined first, and then the search boundary value is determined based on the target row and column values ​​of the target grid, the corresponding grid search data, and the map boundary value, and then the grid search range is determined using the search boundary value. Since the grid is distributed in rows and columns, the map boundary value is represented by row and column values ​​(minimum row value, maximum row value, minimum column value, and maximum column value), and the search boundary value is represented by the minimum grid search row value, the maximum grid search row value, the minimum grid search column value, and the maximum grid search column value.

[0088] In order to ensure that all grids in the grid search range are located in the grid map, the corresponding search boundary value needs to be less than or equal to the corresponding map boundary value. Specifically, there are four cases:

[0089] Case 1: in response to the first difference between the target row value and the grid search quantity being not less than the minimum row value, the first value corresponding to the first difference is used as the minimum grid search row value, otherwise, the minimum row value is used as the minimum grid search row value.

[0090] Case 2: in response to the first summation result of the target row value and the grid search quantity being not greater than the maximum row value, the second value corresponding to the first summation result is used as the maximum grid search row value; otherwise, the maximum row value is used as the maximum grid search row value.

[0091] Case 3: in response to the second difference between the target column value and the grid search quantity being not less than the minimum column value, the third value corresponding to the second difference is used as the minimum grid search column value; otherwise, the minimum column value is used as the minimum grid search column value.

[0092] Case 4: in response to the second summation result of the target column value and the grid search quantity being not greater than the maximum column value, the fourth value corresponding to the second summation result is used as the maximum grid search column value; otherwise, the maximum column value is used as the maximum grid search column value.

[0093] After determining the above minimum grid search row value, maximum grid search row value, minimum grid search column value and maximum grid search column value, the minimum grid search row value, maximum grid search row value, minimum grid search column value and maximum grid search column value are used as search boundary values ​​to construct the grid search range. The above minimum row value, maximum row value, minimum column value and maximum column value are map boundary values ​​of the grid map respectively.

[0094] For example, the resolution of the constructed raster map is 0.5m×0.5m, and the row×column is 200×120, where the row values ​​are (0-199) from bottom to top and the column values ​​are (0-119) from right to left. Then the map boundary value of the raster map is the minimum row value of 0, the maximum row value of 199, the minimum column value of 0, and the maximum column value of 119.

[0095] Further, firstly, based on the preset search range value (for example, 2.5 meters) and the resolution of the grid map of 0.5m×0.5m, the number of grid searches is determined to be 5; then, based on the target row value, target column value and the number of grid searches of 5 for the target grid, each search boundary value is determined. Specifically:

[0096] If the first difference between the target row value and the grid search quantity 5 is not less than the minimum row value 0, the first value corresponding to the first difference is used as the minimum grid search row value; otherwise, if the first difference is less than the minimum row value 0, the minimum row value 0 is determined as the minimum grid search row value. For example, if the target row value is 20, the first difference is 15 (greater than the minimum row value 0), and the corresponding minimum grid search row value is 15.

[0097] If the first sum of the target row value and the grid search quantity 5 is not greater than the maximum row value 199, the second value corresponding to the first sum is used as the maximum grid search row value. Otherwise, if the first sum is greater than 199, the maximum row value 199 is used as the maximum grid search row value. Continuing with the above example, if the target row value is 20 and the first sum is 25 (less than 199), the maximum grid search row value is 25.

[0098] If the second difference between the target column value and the grid search quantity 5 is not less than the minimum column value 0, the third value corresponding to the second difference is used as the minimum grid search column value, otherwise, the minimum column value is used as the minimum grid search column value. For example, if the target column value is 115, the second difference is 110 (not less than the minimum column value 0), then the minimum grid search column value is 110.

[0099] If the second sum of the target column value and the grid search quantity 5 is not greater than the maximum column value 119, the fourth value corresponding to the second sum is used as the maximum grid search column value. Otherwise, the maximum column value is used as the maximum grid search column value. Continuing with the above example, the target column value is 115, and the second sum is 120 (greater than the maximum column value 119), so the maximum grid search column value is the maximum column value 119.

[0100] Based on the above examples, the minimum grid search row value 15, the maximum grid search row value 25, the minimum grid search column value 110 and the maximum grid search column value 119 are used as the search boundary values ​​of the grid search range, that is, the grid search range corresponding to the target grid (row and column values ​​are 20 and 115 respectively) is the 15th to 25th rows and the 110th to 119th columns.

[0101] It should be noted that the points within the grid search range corresponding to each first roadside point are all considered to be points of the same category as the first roadside point and belong to the same target roadside point set, and the points within the grid search range corresponding to each second roadside point are all considered to be points of the same category as the second roadside point and belong to the same target roadside point set.

[0102] It should also be noted that the above method of clustering the roadside points by assigning all detected roadside points to the grid map can directly characterize the distance between the roadside points by the position of the grid in the grid map. Compared with the method of clustering using the distance between each roadside point, the above method of clustering based on the position of the grid where the roadside point is located in the grid map reduces the amount of calculation of the distance between the roadside points. In the search process, the detected roadside points within the grid search range are directly regarded as the same roadside point. A small number of detected roadside points can be used to quickly determine the roadside points of the same category, so the clustering efficiency of the roadside points can be significantly improved.

[0103] S13: Based on the first fused roadside at the historical moment and a plurality of first roadsides at the current moment, a second fused roadside at the current moment is obtained.

[0104] Step S13 is to track the roadside lines at different times in association. The number of the first fused roadside line is at least one.

[0105] Among them, as the vehicle travels, the vehicle coordinate system moves, and the actual roadside lines on the road have relative motion (including rotational motion and translational motion) relative to the vehicle. Before associating the roadside lines at different times, the roadside lines at different times should be expressed in the vehicle coordinate system at the same time (current time).

[0106] Exemplarily, a certain number of curb points are extracted from the first fused curb line, and the vehicle's driving data at historical moments are used to predict the coordinates of each extracted curb point in the vehicle coordinate system at the current moment, so as to obtain the coordinates of the predicted curb points corresponding to each extracted curb point, and then the predicted curb points are used for curve fitting to obtain the predicted curb lines corresponding to each first fused curb line. For ease of description, the predicted curb line corresponding to the first fused curb line at the current moment is referred to as the third fused curb line. Among them, the first fused curb line and the third fused curb line both belong to the tracking curb lines to be tracked at the current moment.

[0107] Among them, a certain number of roadside points can be randomly extracted from the first fused roadside line, or a certain number of roadside points can be extracted from the first fused roadside line at the same distance; of course, a certain number of roadside points can also be extracted according to a certain strategy; for example, distinguishing multiple sections, using different interval distances to take points in each section, etc.

[0108] Among them, the vehicle's driving data at historical moments is used to predict the coordinates of each extracted curb point in the vehicle coordinate system at the current moment. The following formula can be used as a reference:

[0109]

[0110] Where, X k ,y k They respectively represent the ordinate and abscissa of the curb point at time k (current moment) in the vehicle coordinate system (the front of the vehicle is the positive direction of the x-axis, and the left is the positive direction of the y-axis); A represents the rotation transformation matrix of the vehicle coordinate system; B represents the translation transformation matrix of the vehicle coordinate system; ΔX and Δy respectively represent the longitudinal displacement and lateral displacement of the vehicle from time k-1 (historical moment) to time k.

[0111] Further:

[0112]

[0113]

[0114]

[0115]

[0116] Where V k-1 represents the vehicle speed at time k-1, T k-1，k represents the time from time k-1 to time k, and θ represents the angular velocity of the vehicle around the vertical direction, where counterclockwise is positive.

[0117] Furthermore, after determining that each first fused roadside line corresponds to the third fused roadside line at the current moment, each third fused roadside line is associated with a number of first roadside lines determined by multiple target roadside point sets to determine each roadside association pair in which the association is successful; then, the third fused roadside line and the first roadside line in each roadside association pair are fused to obtain each second fused roadside line at the current moment.

[0118] Specifically, each third fused roadside is associated with a number of first roadsides, and a method for determining the successfully associated roadside association pairs can be referred to below. Figure 4 Related description of the illustrated embodiment.

[0119] The above scheme performs rasterization processing on multiple detected curb points of the road at the current moment to obtain a corresponding grid map. The detected curb points can be placed in the grid map, and then clustered using the position of the grid where each detected curb point is located in the grid map to obtain multiple target curb point sets for determining several roadside lines at the current moment. Since the present application can directly use the position of the grid for clustering without considering the position of each detected curb point in the grid, compared to the method that needs to use the position of each point for clustering, the method of clustering using the grid position in the present application is conducive to improving the efficiency of clustering to obtain multiple target curb point sets, that is, it can improve the efficiency of determining the roadside line, and thus is conducive to improving the subsequent curb tracking efficiency.

[0120] See also Figure 4 , Figure 4 1 is a flow chart of an embodiment of determining the association pairs of each roadside provided by the present application. In this embodiment, the above-mentioned association of each third fused roadside with a plurality of first roadsides to determine the association pairs of each roadside that are successfully associated includes:

[0121] S41: Obtain lateral errors between each third fused roadside line and each first roadside line.

[0122] In this embodiment, the lateral error between the third fused roadside line and the first roadside line refers to the lateral error within the effective length of the two roadside lines. Among them, each roadside line is not infinitely extended, and has a corresponding starting point and end point. The effective length refers to the length between the starting point and the end point of the corresponding roadside line. The specific starting point and end point of the roadside line can be determined according to the position of the detected roadside point. Further, the lateral error is used to represent the distance relationship between the third fused roadside line and the first roadside line, which can be but not limited to the average lateral error, and can also be a mean square error, a root mean square error or a standard error, etc.

[0123] S42: Associating the roadside lines based on the lateral errors to obtain successfully associated roadside line association pairs.

[0124] In one embodiment, step S42 further includes the following steps:

[0125] First, based on the lateral errors between each third fused pathline and each first pathline, a cost matrix for characterizing each lateral error is constructed.

[0126] Exemplarily, one of the rows and columns of the cost matrix represents the first roadside line, and the other represents the third fused roadside line, and the elements in the matrix represent the lateral error between the first roadside line and the third fused roadside line.

[0127] Second, roadside line pairing is performed based on the cost matrix, and a roadside line target matching combination is selected from a number of roadside line candidate matching combinations; each roadside line candidate matching combination is a pairing combination including each third fused roadside line and each first roadside line.

[0128] Exemplarily, possible pairing combinations (i.e., several roadside candidate matching combinations) are constructed based on each third fused roadside and each first roadside, wherein each roadside candidate matching combination is a pairing combination including each third fused roadside and each first roadside.

[0129] Then, a roadside target matching combination is selected from the plurality of roadside candidate matching combinations; wherein the sum of the first errors of all roadside matching pairs in the roadside target matching combination is less than the sum of the second errors of all roadside matching pairs in each roadside candidate matching combination; the sum of errors is the sum of the lateral errors corresponding to all roadside matching pairs. That is, after constructing a plurality of roadside candidate matching combinations, the sum of the lateral errors of all roadside matching pairs in each roadside candidate matching combination (the sum of the second errors) is calculated, and then the minimum sum of errors (i.e., the sum of the first errors) is determined from each of the second error sums, and the roadside candidate matching combination corresponding to the minimum sum of errors is used as the roadside target matching combination.

[0130] In a specific implementation, a matching algorithm can be used to perform the above-mentioned roadside matching so that the sum of the lateral errors of all roadside matching pairs in the obtained roadside target matching combination is minimized. For example, Hungarian matching can be used. This matching algorithm is a classic linear programming method used to solve a special type of cost problem, that is, in a two-dimensional cost matrix (that is, the above-mentioned cost matrix for lateral errors), find a pairing method that minimizes the total matching cost while ensuring that each element is matched exactly once. The algorithm gradually constructs an allocation table and uses a series of exchange operations to convert the cost matrix into a "completely matched" state, that is, each row and column has only one non-zero element. At the same time, by comparing the current allocation and the optimal allocation, an exchange is found that minimizes the cost. Corresponding to this, the sum of the lateral errors of the first roadside and the third fused roadside after matching is minimized.

[0131] Third, from the roadside matching pairs in the roadside target matching combination, select the roadside associated pairs that meet the preset requirements; the lateral error between the third fused roadside line in each roadside associated pair and the corresponding first roadside line is less than the preset threshold.

[0132] The above-mentioned roadside pairing based on the cost matrix is ​​only for determining the best pairing combination (roadside target matching combination) between multiple third fused roadsides and multiple first roadsides, but the roadside matching pairs in the best matching combination may not directly correspond to the same roadside on the current road. Therefore, in order to accurately track the roadside subsequently, it is necessary to screen out the matching pairs that truly correspond to the same roadside on the current road from the roadside matching pairs in the roadside target matching combination, that is, to select the roadside association pairs that meet the preset requirements.

[0133] In one embodiment, if the lateral error between the third fused roadside line and the corresponding first roadside line in the roadside line matching pair is less than a preset threshold, the roadside line matching pair is considered to be a matching pair that truly corresponds to the same roadside on the current road, that is, a roadside line association pair. The specific preset threshold can be set according to the specific application scenario and road conditions.

[0134] After the above-mentioned roadside line association is performed, there may still be first roadside lines or third fused roadside lines that have not been successfully associated. Among them, the reason for the existence of first roadside lines that have not been successfully associated may be: the number of first roadside lines is greater than the number of third fused roadside lines, or the lateral errors between the first roadside lines and each third fused roadside line are greater than the preset threshold. Similarly, the reason for the existence of third fused roadside lines that have not been successfully associated may be: the number of third fused roadside lines is greater than the number of first roadside lines, or the lateral errors between the third fused roadside lines and each first roadside line are greater than the preset threshold.

[0135] Therefore, in some embodiments, after the roadside lines are associated based on the lateral errors to obtain the successfully associated roadside line association pairs, at least one of the following steps is further included:

[0136] First, determine whether there is any third fusion route that has not been successfully associated among the third fusion routes; in response to the existence of a third fusion route that has not been successfully associated, use the third fusion route that has not been successfully associated as the fourth fusion route; in response to the absence of a first route that has been successfully associated with the fourth fusion route in the first preset number of associations in the future, delete the fourth fusion route.

[0137] It should be noted that the third fused roadside is used as a tracking roadside for roadside tracking. If there is a third fused roadside that has not been successfully associated (for the sake of distinction, it is called the fourth fused roadside), it may be because the fourth fused roadside has disappeared (for example, in the case of a turn at an intersection, the original roadside will disappear), or it may be due to a sensor detection error, etc. Therefore, in order to avoid the situation where the roadside still exists but is mistakenly believed to have disappeared, the fourth fused roadside can be retained at the current moment and associated with the first roadside in the future. If multiple associations (the first preset number of associations) have not been successfully associated with the corresponding first roadside, the fourth fused roadside is considered to be a roadside that has disappeared, and the fourth fused roadside is deleted accordingly, and it will not be used as a tracking roadside for subsequent tracking. The specific first preset number of associations can be determined based on actual scene experience.

[0138] Second, determine whether there is a first roadside line that has not been successfully associated among the first roadside lines; in response to the existence of a first roadside line that has not been successfully associated, use the first roadside line that has not been successfully associated as a candidate tracking roadside line; in response to the existence of a first roadside line associated with the candidate tracking roadside line in the future second preset number of associations, determine the candidate tracking roadside line as a tracking roadside line.

[0139] It should be noted that if there is a first roadside line that has not been successfully associated, it may be because the first roadside line is a newly added roadside line on the road (for example, in a turning scenario at an intersection), or there may be an error in the roadside line detection. If it is a newly added roadside line of a real road, then the roadside line will be continuously tracked as a tracking roadside line. Therefore, when it is determined that there is a first roadside line that has not been successfully associated, the first roadside line that has not been successfully associated will be first used as a candidate tracking roadside line and associated with the future first roadside line. If there are multiple successful associations with the first roadside line in the future (a second preset number of associations), the candidate tracking roadside line is considered to be a roadside line on the real road, and it is determined as a tracking roadside line for subsequent roadside tracking.

[0140] Among them, the above-mentioned first fused roadside line and the third fused roadside line both belong to tracking roadside lines. The difference is that the first fused roadside line is represented in the vehicle coordinate system corresponding to the historical moment, and the third fused roadside line is represented in the vehicle coordinate system corresponding to the current moment.

[0141] In some embodiments, the multiple target roadside point sets obtained by clustering in step S12 are used to determine a number of first roadside lines at the current moment, wherein a first roadside line is used to represent a roadside on the road.

[0142] In one embodiment, see Figure 5 , Figure 5 This is a flow chart of an embodiment of determining several first routes at the current moment provided by the present application. This embodiment includes:

[0143] S51: Obtain second roadside lines determined based on each target roadside point set.

[0144] In this embodiment, the first curve fitting can be performed using each target roadside point set to obtain the second roadside line corresponding to each target roadside point set. That is, a second roadside line is obtained by fitting a target roadside point set. The second roadside line may correspond to a roadside on a real road, or may be a section of a roadside on a real road. For example, a roadside is blocked by an obstacle, and what is seen from the perspective of the vehicle is two roadside lines, but in fact, these two roadside lines are different roadside sections of the same roadside on the road.

[0145] In one embodiment, the least square method may be used to perform first curve fitting based on each target roadside point set to obtain a fitted second roadside line.

[0146] In one implementation, all detected curb points in a target curb point set may be used as interior points, and curve fitting may be performed using the least squares method to obtain a corresponding second curb line.

[0147] In another embodiment, some roadside points may be randomly selected from the target roadside point set and assumed as interior points on the curve, and then the randomly selected interior points are fitted using the least squares method to obtain the initial curve equation corresponding to the initial curve. Exemplarily, it can be expressed as follows:

[0148] y=C 3 X 3 +C 2 X 2 +C 1 X+C 0

[0149] In the formula, y is the horizontal coordinate of the point on the curve, x is the vertical coordinate of the point on the curve, C 3 , C 2 , C 1and C 0 They are the cubic term coefficient, quadratic term coefficient, linear term coefficient and constant term coefficient respectively. It should be noted that unless otherwise specified in this application, the vehicle coordinate system is a coordinate system constructed with the forward direction as the X-axis and the lateral direction as the Y-axis.

[0150] Then, the distance between the remaining curb points in the target curb point set and the initial curve is calculated. If the distance is less than the preset distance threshold, the corresponding curb point is regarded as an inlier. Specifically, the ordinate (x) of a curb point can be substituted into the initial curve equation to obtain the corresponding abscissa (y). If the difference between the calculated abscissa and the actual abscissa of the curb point is less than the preset distance threshold, the curb point is regarded as an inlier. The number of inliers in the target curb point set is counted. If the number of inliers is greater than or equal to the preset inlier threshold, it means that the initial curve obtained by fitting is better. At this time, the inliers in the currently determined target curb point set are retained.

[0151] Furthermore, if the number of inliers is less than the preset inlier number threshold, repeat the above steps (randomly select some of the roadside points from the target roadside point set as the inliers on the curve, and subsequent steps) until the number of inliers counted is greater than or equal to the preset inlier number threshold, retaining each inlier in the currently determined target roadside point set, or until the number of repetitions reaches the preset number, retaining those inliers with the largest number of inliers in the preset number. The specific preset inlier number threshold and the preset number are determined based on experience.

[0152] Finally, the curve fitting is performed again based on the retained interior points to obtain the second roadside line corresponding to the target roadside point set.

[0153] Furthermore, in one embodiment, the inner points in each target roadside point set may be sorted, so as to use the sorted inner points to determine the starting roadside point and the ending roadside point on the second roadside line. The starting roadside point and the ending roadside point are determined according to the vehicle's driving direction. Specifically, compared with the ending roadside point, the starting roadside point is a roadside point that is farther from the driving direction.

[0154] The above-mentioned least square method is an optional curve fitting method in this application. The specific curve fitting method can be selected according to actual needs.

[0155] S52: Determine target roadside line pairs satisfying a splicing condition from the second roadside lines.

[0156] In this embodiment, the target roadside line pairs that meet the splicing conditions are determined to be target roadside line pairs corresponding to the same roadside on the real road. In simple terms, the two second roadside lines in each target roadside line pair are roadside lines corresponding to different roadside sections of the same roadside on the real road.

[0157] In one embodiment, candidate roadside line pairs to be matched may be selected from the second roadside lines. The two second roadside lines in each candidate roadside line pair may be roadside lines on different roadside sections on the same roadside of the road, or may be two roadside lines corresponding to different roadsides on the road.

[0158] For each candidate roadside line pair, based on the first position difference between the two second roadside lines in the candidate roadside line pair, determine whether the two second roadside lines meet the first splicing condition; if the first splicing condition is met, determine that the two second roadside lines meet the splicing condition, and use the two second roadside lines that meet the splicing condition as the two second roadside lines in a target roadside line pair.

[0159] The first position difference includes at least one of a lateral position difference and an angle difference; the first splicing condition includes that the lateral position difference is less than a lateral threshold, and / or the angle difference is less than an angle threshold. Specifically, the lateral threshold and the angle threshold are determined based on actual experience.

[0160] In one implementation scenario, if the lateral position difference between the two second roadside lines is less than a lateral threshold, or the angle difference is less than an angle threshold, it is determined that the two second roadside lines meet the first splicing condition.

[0161] In another implementation scenario, in order to avoid mistakenly splicing two second roadside lines belonging to different roadsides into one first roadside line, it is required that the lateral position difference between the two second roadside lines is less than the lateral threshold and the angle difference is less than the angle threshold, only then is it determined that the two second roadside lines meet the first splicing condition.

[0162] In one embodiment, determining the lateral difference between the two second road edges comprises the following steps:

[0163] First, one of the two second roadside lines is used as a first target roadside line, the other is used as a second target roadside line, and a target point is selected from the first target roadside line.

[0164] Second, obtaining an associated target point of the target point, where the associated target point is obtained by extending the second target roadside line, and the longitudinal position difference between the target point and the associated target point is less than a preset difference threshold;

[0165] Third, the lateral position difference between the target point and the associated target point is used as the lateral position difference between the two second road lines.

[0166] For example, see Figure 6 , Figure 6 This is a schematic diagram of an embodiment of determining a target road along the route of the present application. Figure 6 As shown, the roadside line (second roadside line) l 1Heluanxian (Second Luanxian) 2 One of them is used as the first target roadside line, and the other is used as the second target roadside line. 1 As the first target roadside line, the roadside line l 2 As the second target roadside line, then from the roadside line l 1 Select the target point (P 1 );Then get the target point P 1 The associated target point (for example, Q 1 ), where the associated target point Q 1 For the second target roadside line (roadside line l 2 ) is extended and associated with the target point Q 1 and the target point P 1 The longitudinal position difference between them is less than the preset difference threshold; determine the associated target point Q 1 and the target point P 1 and take it as the lateral position difference between the two second road lines.

[0167] It should be noted that the above is only used as an example and cannot limit the present application. For example, any position point on the first target roadside line can be used as the target point, or the detected roadside point on the first target roadside line close to the second target roadside line can be used as the target point. The number of target points can be one or more, and correspondingly, the number of associated target points can also be one or more. Of course, in order to enhance the comparability between the two roadside lines, the longitudinal coordinates of the associated target point and the target point can be set to be the same, such as Figure 6 As shown, the coordinates in the x direction are the same.

[0168] In a specific embodiment, the target point is the starting point or the ending point of the second roadside line. 1 The starting point of the roadside is taken as the target point P 1 .

[0169] In one embodiment, determining the angle difference between the two second road edges comprises the following steps:

[0170] First, obtain the slope representation value of each of the two second road lines.

[0171] Second, the difference between the slope characterization values ​​corresponding to the two second roadside lines is used as the angle difference between the two second roadside lines.

[0172] In one embodiment, the slope characterization value corresponding to the first target roadside line is the first slope value of the curve equation corresponding to the first target roadside line at the target point, and the slope characterization value corresponding to the second target roadside line is the second slope value of the curve equation corresponding to the second target roadside line at the associated target point. The curve equation corresponding to each second roadside line is obtained by curve fitting the corresponding target roadside point set.

[0173] Optionally, the curve equation can be differentiated to obtain the derivative equation of the curve equation (which is the slope equation), and then the target point and the associated target point are respectively substituted into the corresponding slope equation to obtain the slope representation value of each of the two second roadside lines. 1 Substitute the roadside line l 1 The corresponding slope equation will be associated with the target point Q 1 Substitute the roadside line l 2 The corresponding slope equations are used to obtain the corresponding slope characterization values.

[0174] It should be noted that the curb is divided into two sections in at least one of the following two situations.

[0175] Case 1: Intersection scenario. In the intersection scenario, two sections of the roadside line will be discontinuous. However, in this case, the two sections of the roadside line are discontinuous because the two sections of the roadside before and after the intersection belong to different roadsides. The purpose of this application to determine the target roadside line pair that meets the splicing conditions is to splice the second roadside lines corresponding to the two sections of the roadside that should have been the same roadside into a complete first roadside line. In the intersection scenario, the two sections of the roadside line are discontinuous because the intersection divides the two roadsides, not because a roadside is divided into two parts. In this scenario, if the two second roadside lines corresponding to the two sections of the roadside before and after the intersection are spliced ​​into a first roadside line, the vehicle control end will mistakenly regard the intersection as a normal road, which will affect the planning of future vehicle driving trajectories.

[0176] Case 2: Obstacle occlusion causes a curb to be divided into two sections. From the perspective of the vehicle, if there is an obstacle, it will cause the obstacle to block part of the curb, resulting in a situation where the curb is divided into two sections. In this case, the two curbs seen by the vehicle are actually one curb on the road. Therefore, in order to accurately associate the first curb and the first fused curb in the future, the two discontinuous (segmented) second curbs caused by obstacle occlusion can be spliced ​​into one first curb.

[0177] Of course, there may be a scene where both the intersection and the scene blocked by obstacles exist at the same time. Therefore, if you want to accurately distinguish the above scenes, you need to set a second splicing condition that can distinguish the above scenes.

[0178] Therefore, in one embodiment, when it is determined that the two second roadside lines in the candidate roadside line pair meet the first splicing condition, it is also necessary to further determine whether the two second roadside lines in the candidate roadside line pair meet the second splicing condition based on the obstacle detection result at the current moment. If the second splicing condition is met, it is determined that the two second roadside lines meet the splicing condition, and the two second roadside lines that meet the splicing condition are used as the two second roadside lines in a target roadside line pair.

[0179] Specifically, the obstacle detection result at the current moment includes one of a first detection result that no obstacle exists and a second detection result that an obstacle exists, that is, the obstacle detection result includes two results: an obstacle is detected or an obstacle is not detected.

[0180] In one embodiment, based on the obstacle detection result at the current moment, determining whether two second roadside lines in the candidate roadside line pair meet the second splicing condition includes the following situations:

[0181] Case 1: If the detection result is that there is no obstacle, it is determined that the two second roads do not meet the second splicing condition.

[0182] Case 2: If the detection result is the second detection result indicating the presence of an obstacle, the second position difference between the occlusion area corresponding to the obstacle and the two second roadside lines is further determined, and based on the second position difference, it is determined whether the two second roadside lines meet the second splicing condition.

[0183] In one embodiment, see Figure 7 , Figure 7 : is a flow chart of an embodiment of determining the second position difference provided by the present application. In this embodiment, determining the second position difference between the shielding area corresponding to the obstacle and the two second road lines includes:

[0184] S71: Determine first intersection points between two second road lines or corresponding extension lines and the shielding area respectively.

[0185] In one embodiment, the second detection result includes a first position of an obstacle detection frame, and the obstacle detection frame is detected by using an existing related detection model or algorithm. Since it is a prior art, it will not be described in detail here.

[0186] The occlusion area corresponding to the obstacle is determined by using the first position of the obstacle detection frame, or by using the second positions corresponding to the four corner points of the obstacle detection frame, and the second positions corresponding to the corner points are determined based on the first position of the obstacle detection frame.

[0187] Exemplarily, the first position of the obstacle detection frame includes the coordinates of the center point of the detection frame, the size information of the frame, and the orientation angle (the angle with the positive direction of the X-axis, which is positive in the counterclockwise direction). The center point coordinates are, for example, O'(x, y), and the size information of the frame is, for example, the length, width, and height of the frame (l, w, h). Based on the first position of the obstacle detection frame, the second position corresponding to each corner point (A, B, C, D) in the obstacle detection frame can be determined. The relevant calculation formula is as follows:

[0188]

[0189] Among them, X A , X B , X C , X D are the ordinates of the four corner points A, B, C, and D, respectively, A ,y B ,y C ,y D are the horizontal coordinates of the four corner points A, B, C, and D, θ is the angle between the orientation of the detection frame and the positive direction of the X-axis, counterclockwise is positive, l is the length of the detection frame, and w is the width of the detection frame.

[0190] Of course, in other embodiments, the relevant detection algorithm may also be directly used to determine the second position corresponding to each corner point on the obstacle detection frame.

[0191] In one embodiment, see Figure 8 , Figure 8 1 is a flow chart of an embodiment of determining an occlusion area of ​​an obstacle provided by the present application. In this embodiment, determining the occlusion area of ​​an obstacle includes:

[0192] S81: In response to the obstacle being located on one side of the vehicle's forward direction, based on the second positions of the preset position points and the corner points on the vehicle at the current moment, connecting lines between the preset position points and the corner points are constructed.

[0193] S82: Obtain the angles between each connecting line and the vehicle's forward direction, and use the connecting line corresponding to the smallest of the angles as the first connecting line, and use the connecting line corresponding to the largest of the angles as the second connecting line.

[0194] S83: Using the area between the first connection line and the second connection line as the shielding area.

[0195] In this embodiment, considering that there may be an obstacle directly in front of the vehicle on the road, in this case, there will not be a situation where a complete first roadside line that should have been the same roadside is blocked by the obstacle directly in front of the vehicle and is thus divided into two second roadside lines. Figure 8Before determining the blocked area in the embodiment shown, it is first determined whether the obstacle is located directly in front of the vehicle. If the obstacle is located directly in front of the vehicle, the blocking area is not determined. Figure 8 In the steps of the embodiment shown in the figure, if the obstacle is located on one side of the vehicle's forward direction, it means that the obstacle is not directly in front of the vehicle. In this case, the following steps are performed: Figure 8 Steps of the illustrated embodiment.

[0196] Among them, the method of determining whether the obstacle is located on one side of the vehicle's forward direction can be, but is not limited to, determined by the positions of the corner points of the obstacle. It can also be determined in other ways, such as by the lateral and longitudinal differences between the center point of the obstacle and the vehicle's forward direction, or directly detected using a related detection model, etc.

[0197] Among them, whether the obstacle is located on one side of the vehicle's forward direction is determined by the position of each corner point of the obstacle. For example, it is determined whether each corner point of the obstacle is on the same side of the vehicle's forward direction. If they are on the same side, it means that the obstacle is located on one side of the vehicle's forward direction. In this embodiment, the vehicle's forward direction is, for example, Figure 6 The direction of the x-axis is shown.

[0198] Please continue reading Figure 6 , based on the second position of the preset position point O and each corner point (A, B, C, D) on the vehicle at the current moment, connect the preset position point O and each corner point (A, B, C, D) to construct a connection line between the preset position point and each corner point, then obtain the angle between each connection line and the vehicle's forward direction, and use the connection line corresponding to the smallest of the angles as the first connection line, and use the connection line corresponding to the largest of the angles as the second connection line; then use the area between the first connection line and the second connection line as the blocked area. Among them, the blocked area includes the blocked roadside line l 1 And along the road 2 The roadside area between the preset position points and the corner points can be constructed by connecting the preset position points with the corner points (for example, OA, OB, OC, OD), or by extending the corresponding line segments, that is, by extending OA, OB, OC, OD. Figure 6 As shown, the first connecting line is the connecting line where the line segment OA is located, and the second connecting line is the connecting line where the line segment OC is located. In this embodiment, the preset position point O can be the origin of the vehicle coordinate system, or it can be the installation position point of the image acquisition device on the vehicle.

[0199] In one embodiment, the first intersection points of the two second roadside lines or corresponding extension lines with the occlusion area determined in step S71 include: the second intersection points of the two second roadside lines or corresponding extension lines with the first connecting line, and the third intersection points of the two second roadside lines or corresponding extension lines with the second connecting line.

[0200] For example, Figure 6 As shown, along the road 1 And along the road 2 The second intersection points of the extension lines of the first connecting line are P1 and Q1, and the roadside line l 1 The extension line and roadside line 2 The third intersection points with the second connecting line are P4 and Q4 respectively.

[0201] S72: Taking the longitudinal position difference between the target roadside point on each second roadside line and the first intersection point of the corresponding second roadside line and the shielding area as the second position difference.

[0202] It should be noted that the second roadside line includes a starting roadside point and an ending roadside point, and the starting roadside point and the ending roadside point are determined based on the position sorting of each detected roadside point in the target roadside point set corresponding to the second roadside line. Exemplarily, the starting roadside point and the ending roadside point can be determined based on the position sorting of the inner point in the target roadside point set corresponding to the second roadside line. The specific method of determining the inner point can refer to the description of the curve fitting part above, and no further details will be given here.

[0203] In one embodiment, the target roadside point on the front roadside line of the two second roadside lines is the starting roadside point on the front roadside line, and the target roadside point on the rear roadside line is the ending roadside point on the rear roadside line. Figure 6 , Figure 6 Along Zhongqian Road 1 The target roadside point is the roadside line l 1 The starting point on the curb, the rear curb line l 2 The target roadside point on the roadside line l 2 The destination curb point on the road, in simple terms, the target curb point on each second curb line is a position point on the corresponding curb line close to the blocked area.

[0204] Of course, in other embodiments, the target curb point on the front curb line of the two second curb lines can also be a position point close to the starting curb point on the front curb line, and the target curb point on the rear curb line can be a position point close to the ending curb point on the rear curb line. The specific positions and numbers of the target curb points can be selected according to actual needs.

[0205] Further, after determining the target curb point on each second curb line, the longitudinal position difference between the target curb point on each second curb line and the first intersection point of the corresponding second curb line and the shielding area is determined.

[0206] In one embodiment, the longitudinal position differences between the starting curb point on the front curb line and the two second intersection points can be determined, as well as the longitudinal position differences between the ending curb point on the rear curb line and the two third intersection points, and the four longitudinal position differences obtained are used as the second position differences.

[0207] In another embodiment, a second intersection point farther from the vehicle can be selected from the two second intersection points as the first intersection point to be compared, and a third intersection point closer to the vehicle can be selected from the two third intersection points as the second intersection point to be compared. Then, a first longitudinal position difference between the starting curb point on the front curb line and the first intersection point to be compared is determined, and a second longitudinal position difference between the ending curb point on the rear curb line and the second intersection point to be compared is determined, and the first longitudinal position difference and the second longitudinal position difference are used as the second position difference.

[0208] For example, Figure 6 As shown, along the road 1 And along the road 2 The second intersection points of the extension lines of the first connecting line are P1 and Q1, and the roadside line l 1 The extension line and roadside line 2 The third intersection points with the second connecting line are P4 and Q4 respectively, and the roadside line l 1 The first longitudinal position difference between the starting curb point and the second intersection point P1, and the curb line l 2 The second longitudinal position difference between the end curb point and the third intersection point Q4 is taken as the second position difference.

[0209] The second splicing condition includes: each longitudinal position difference is not greater than the longitudinal difference threshold. That is, if the longitudinal position difference between the target roadside point and the corresponding intersection on each second roadside line is less than or equal to the longitudinal difference threshold, it is determined that the two second roadside lines in the candidate roadside line pair meet the second splicing condition.

[0210] S53: Splicing two second roadside lines in each target roadside line pair to obtain a corresponding spliced ​​roadside line.

[0211] In this embodiment, splicing two second roadside lines to obtain a corresponding spliced ​​roadside line refers to supplementing the roadside lines of two discontinuous second roadside lines so as to splice the two discontinuous second roadside lines into a continuous spliced ​​roadside line.

[0212] In one embodiment, for each target roadside line pair, the target roadside point set corresponding to each second roadside line in the target roadside line pair may be first merged to obtain a merged roadside point set, and then a second curve fitting is performed on the merged roadside point set to obtain a spliced ​​roadside line. The specific second curve fitting method can refer to the first curve fitting method mentioned above, which will not be described in detail here.

[0213] Of course, in one embodiment, only the inner points in the target roadside point set corresponding to each second roadside line may be merged to obtain a merged inner point set, and then a second curve fitting is performed on the merged inner point set to obtain a spliced ​​roadside line. The specific method of determining the inner points can refer to the introduction of the first curve fitting part above.

[0214] S54: taking each spliced ​​roadside line as a first roadside line, and taking each remaining second roadside line that does not meet the splicing condition as a first roadside line, to obtain a plurality of first roadside lines.

[0215] In summary, this embodiment splices two discontinuous second roadside lines that can be spliced ​​into one first roadside line, and each second roadside line that cannot be spliced ​​is used as a first roadside line separately, so as to track the road sideline using several first roadside lines and the first fused roadside line at a historical moment.

[0216] See also Fig. 9 , Fig. 9 1 is a schematic diagram of a framework of an electronic device according to an embodiment of the present application. In this embodiment, the electronic device 90 includes a memory 91 and a processor 92 coupled to each other.

[0217] The memory 91 stores program instructions, and the processor 92 is used to execute the program instructions stored in the memory 91 to implement the steps of any of the above-mentioned method implementation methods. In a specific implementation scenario, the electronic device 90 may include, but is not limited to: a microcomputer, a server, and in addition, the electronic device 90 may also include a mobile device such as a laptop computer and a tablet computer, which is not limited here.

[0218] Specifically, the processor 92 is used to control itself and the memory 91 to implement the steps of any of the above-mentioned embodiments. The processor 92 can also be called a CPU (Central Processing Unit). The processor 92 may be an integrated circuit chip with signal processing capabilities. The processor 92 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 92 can be implemented by an integrated circuit chip.

[0219] See also Fig.10 , Fig.10 It is a schematic diagram of the framework of the computer-readable storage medium provided by the present application. The computer-readable storage medium 100 of the embodiment of the present application stores a program instruction 101, and the program instruction 101 is executed to implement the method provided by any embodiment of the above method and any non-conflicting combination. Among them, the program instruction 101 can form a program file and be stored in the above-mentioned computer-readable storage medium 100 in the form of a software product, so that a computer device (which can be a personal computer, a server, or a network device, etc.) executes all or part of the steps of each implementation method of the present application. The aforementioned computer-readable storage medium 100 includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, or a terminal device such as a computer, a server, a mobile phone, and a tablet.

[0220] See also Fig.11 , Fig.11 It is a framework diagram of an embodiment of a curb tracking device provided by the present application. In this embodiment, the curb tracking device 110 includes a processing module 111, a clustering module 112 and a tracking module 113. The processing module 111 is used to perform rasterization processing on multiple detected curb points of the road at the current moment to obtain a corresponding grid map; the clustering module 112 is used to cluster the positions of the grids where each detected curb point is located in the grid map to obtain multiple target curb point sets; the multiple target curb point sets are used to determine several first curb lines at the current moment; the tracking module 113 is used to obtain the second fused curb line at the current moment based on the first fused curb line at the historical moment and several first curb lines at the current moment.

[0221] In some embodiments, the clustering module 112 uses the position of the grid where each detected roadside point is located in the grid map to perform clustering to obtain multiple target roadside point sets, including: selecting an unsearched roadside point from the multiple detected roadside points as the first roadside point; searching for detected roadside points of the same category as the first roadside point from the first grid search range to obtain an initial roadside point set corresponding to the first roadside point; the first grid search range is determined based on the first position of the first grid where the first roadside point is located in the grid map, and is used to represent the grid range to be searched in the grid map; based on the initial roadside point set corresponding to the first roadside point, determine the target roadside point set corresponding to the first roadside point; the target roadside point set includes all detected roadside points of the same category as the first roadside point among all detected roadside points; repeat the above steps until all detected roadside points in the grid map have been searched, and obtain the target roadside point set corresponding to each first roadside point.

[0222] In some embodiments, based on the initial roadside point set corresponding to the first roadside point, determining the target roadside point set corresponding to the first roadside point includes: taking each detected roadside point in the initial roadside point set except the first roadside point as a second roadside point; for each second roadside point, searching for a detected roadside point of the same category as the second roadside point from a second grid search range, and adding the searched detected roadside point to the initial roadside point set; determining the second grid search range based on a second position of the second grid where the second roadside point is located in the grid map; adding each searched detected roadside point to the initial roadside point set; The first roadside point is respectively taken as a new second roadside point, and the step of searching for detected roadside points of the same category as the second roadside point from the second grid search range for each second roadside point is returned, and the searched detected roadside points are added to the initial roadside point set, until there are no detected roadside points of the same category as the second roadside points in the second grid search range, or the number of detected roadside points of the same category as the second roadside points is less than the set threshold, and the last determined initial roadside point set is taken as the target roadside point set corresponding to the first roadside point.

[0223] In some embodiments, the grid map includes a plurality of grids distributed in rows and columns, and the position of the grid where the detected roadside point is located in the grid map includes the row and column values ​​of the grid in the grid map; the detected roadside points of the same category as the first roadside point among all the detected roadside points include: the second roadside points located in the first grid search range corresponding to the first roadside point, and the detected roadside points located in the second grid search range corresponding to the second roadside points; the first grid search range is determined by using the first row and column values ​​of the first grid where the first roadside point is located in the grid map and a preset search range value; the second grid search range is determined based on the second row and column values ​​of the second grid where the second roadside point is located in the grid map and a preset search range value.

[0224] In some embodiments, the preset search range value is smaller than the width of the road; the step of determining the grid search range includes: determining the number of grid searches using the preset search range value and the resolution of the grid map; determining the grid search range based on the target row and column values ​​of the target grid in the grid map and the number of grid searches; wherein the grid search range is any one of a first grid search range and a second grid search range, and in response to the grid search range being the first grid search range, the target grid represents the first grid where the first roadside point is located; in response to the grid search range being the second grid search range, the target grid represents the second grid where the second roadside point is located.

[0225] In some embodiments, the target row and column values ​​include the target row value and the target column value of the target grid in the grid map; based on the target row and column values ​​of the target grid in the grid map and the grid search quantity, determining the grid search range includes: in response to a first difference between the target row value and the grid search quantity being not less than a minimum row value, taking a first numerical value corresponding to the first difference as the minimum grid search row value, otherwise, taking the minimum row value as the minimum grid search row value; in response to a first summation result of the target row value and the grid search quantity being not greater than a maximum row value, taking a second numerical value corresponding to the first summation result as the maximum grid search row value, otherwise, taking the maximum row value as the maximum grid search row value; in response to a first summation result of the target row value and the grid search quantity being not greater than a maximum row value, taking a second numerical value corresponding to the first summation result as the maximum grid search row value, otherwise, taking the maximum row value as the maximum grid search row value; The second difference between the grid search quantities is not less than the minimum column value, and the third value corresponding to the second difference is used as the minimum grid search column value, otherwise, the minimum column value is used as the minimum grid search column value; in response to the second summation result of the target column value and the grid search quantity being not greater than the maximum column value, the fourth value corresponding to the second summation result is used as the maximum grid search column value, otherwise, the maximum column value is used as the maximum grid search column value; the minimum grid search row value, the maximum grid search row value, the minimum grid search column value and the maximum grid search column value are respectively used as search boundary values ​​to construct a grid search range; wherein the minimum row value, the maximum row value, the minimum column value and the maximum column value are respectively map boundary values ​​of the grid map.

[0226] In some embodiments, the roadside tracking device 110 also includes a roadside determination module, which is used to determine several first roadside lines at the current moment, including: obtaining second roadside lines determined based on each target roadside point set; determining each target roadside line pair that meets the splicing condition from each second roadside line; splicing based on the two second roadside lines in each target roadside line pair to obtain a corresponding spliced ​​roadside line; taking each spliced ​​roadside line as a first roadside line, and taking each remaining second roadside line that does not meet the splicing condition in each second roadside line as a first roadside line, to obtain several first roadside lines.

[0227] In some embodiments, determining target roadside line pairs that meet the splicing condition from the second roadside lines includes: selecting candidate roadside line pairs to be matched from the second roadside lines; for each candidate roadside line pair, determining whether the two second roadside lines meet the first splicing condition based on the first position difference between the two second roadside lines in the candidate roadside line pair; in response to satisfying the first splicing condition, determining that the two second roadside lines meet the splicing condition; or, in response to satisfying the first splicing condition, determining whether the two second roadside lines meet the second splicing condition based on the detection result of the obstacle at the current moment; in response to the two second roadside lines satisfying the second splicing condition, determining that the two second roadside lines meet the splicing condition; and using the two second roadside lines that meet the splicing condition as the two second roadside lines in a target roadside line pair.

[0228] In some embodiments, the first position difference includes at least one of a lateral position difference and an angle difference; the first stitching condition includes that the lateral position difference is less than a lateral threshold, and / or the angle difference is less than an angle threshold.

[0229] In some embodiments, determining the lateral difference includes: taking one of the two second roadside lines as the first target roadside line and the other as the second target roadside line, and selecting a target point from the first target roadside line; obtaining an associated target point of the target point, the associated target point being obtained by extending the second target roadside line, and the longitudinal position difference between the target point and the associated target point being less than a preset difference threshold; taking the lateral position difference between the target point and the associated target point as the lateral position difference between the two second roadside lines; determining the angle difference includes: obtaining a slope characterization value of each of the two second roadside lines; and taking the difference between the slope characterization values ​​corresponding to the two second roadside lines as the angle difference between the two second roadside lines.

[0230] In some embodiments, the slope characterization value corresponding to the first target roadside line is the first slope value of the curve equation corresponding to the first target roadside line at the target point, and the slope characterization value corresponding to the second target roadside line is the second slope value of the curve equation corresponding to the second target roadside line at the associated target point; the curve equation corresponding to each second roadside line is obtained by curve fitting the corresponding target roadside point set; and / or, the target point is a detected roadside point on the first target roadside line close to the second target roadside line.

[0231] In some embodiments, the obstacle detection result includes one of a first detection result indicating that there is no obstacle and a second detection result indicating that there is an obstacle; based on the obstacle detection result at the current moment, determining whether the two second roadside lines meet the second splicing condition includes: in response to the detection result being the first detection result, determining that the two second roadside lines do not meet the second splicing condition; in response to the detection result being the second detection result, determining a second position difference between an occlusion area corresponding to the obstacle and the two second roadside lines; based on the second position difference, determining whether the two second roadside lines meet the second splicing condition.

[0232] In some embodiments, the second detection result includes the first position of the obstacle detection frame, and the occlusion area is determined using the first position of the obstacle detection frame; determining the second position difference between the occlusion area corresponding to the obstacle and the two second roadside lines, including: determining the first intersection points of the two second roadside lines or the corresponding extension lines with the occlusion area respectively; taking the longitudinal position difference between the target roadside point on each second roadside line and the first intersection point of the corresponding second roadside line and the occlusion area as the second position difference; the second splicing condition includes: each longitudinal position difference is not greater than the longitudinal difference threshold.

[0233] In some embodiments, the second roadside line includes a starting roadside point and an ending roadside point, and the starting roadside point and the ending roadside point are determined based on the position sorting of each detected roadside point in the target roadside point set corresponding to the second roadside line; the target roadside point on the front roadside line of the two second roadside lines is the starting roadside point on the front roadside line, and the target roadside point on the rear roadside line is the ending roadside point on the rear roadside line.

[0234] In some embodiments, the occlusion area corresponding to the obstacle is determined by using the second positions corresponding to the four corner points of the obstacle detection frame, and the second positions corresponding to the corner points are determined based on the first position of the obstacle detection frame; determining the occlusion area includes: constructing connecting lines between the preset position points and the corner points based on the second positions of the preset position points on the vehicle at the current moment; obtaining the angles between each connecting line and the vehicle's forward direction, and taking the connecting line corresponding to the smallest of the angles as the first connecting line, and taking the connecting line corresponding to the largest of the angles as the second connecting line; taking the area between the first connecting line and the second connecting line as the occlusion area; the first intersection points of the two second roadside lines or the corresponding extension lines with the occlusion area include: the second intersection points of the two second roadside lines or the corresponding extension lines with the first connecting line, and the third intersection points of the two second roadside lines or the corresponding extension lines with the second connecting lines.

[0235] In some embodiments, obtaining a second roadside line determined based on each target roadside point set includes: performing a first curve fitting using each target roadside point set to obtain a second roadside line corresponding to each target roadside point set; and / or, splicing two second roadside lines in each target roadside line pair to obtain a corresponding spliced ​​roadside line, including: for each target roadside line pair, merging the target roadside point sets corresponding to each second roadside line in the target roadside line pair to obtain a merged roadside point set; performing a second curve fitting on the merged roadside point set to obtain a spliced ​​roadside line.

[0236] In some embodiments, the number of first fused roadside lines is at least one; the tracking module 113 obtains the second fused roadside line at the current moment based on the first fused roadside lines at the historical moment and several first roadside lines at the current moment, including: based on the driving data of the vehicle at the historical moment, predicting that each first fused roadside line at the historical moment corresponds to the third fused roadside line at the current moment; associating each third fused roadside line with several first roadside lines to determine each roadside association pair that is successfully associated; for each roadside association pair, the third fused roadside line and the first roadside line in the roadside association pair are fused to obtain the second fused roadside line at the current moment.

[0237] In some embodiments, each third fused roadside line is associated with a number of first roadside lines to determine each successfully associated roadside line association pair, including: obtaining the lateral error between each third fused roadside line and each first roadside line respectively; and associating the roadside lines based on each lateral error to obtain each successfully associated roadside line association pair.

[0238] In some embodiments, roadside lines are associated based on each lateral error to obtain each successfully associated roadside line association pair, including: based on the lateral errors between each third fused roadside line and each first roadside line, a cost matrix for characterizing each lateral error is constructed; roadside lines are paired based on the cost matrix, and a roadside line target matching combination is selected from a number of roadside line candidate matching combinations; each roadside line candidate matching combination is a pairing combination including each third fused roadside line and each first roadside line; from each roadside line matching pair in the roadside line target matching combination, each roadside line association pair that meets preset requirements is selected; the lateral error between the third fused roadside line in each roadside line association pair and the corresponding first roadside line is less than a preset threshold.

[0239] In some embodiments, the sum of the first errors of all roadside matching pairs in the roadside target matching combination is less than the sum of the second errors of all roadside matching pairs in each roadside candidate matching combination; the sum of the errors is the sum of the lateral errors corresponding to all roadside matching pairs; after the roadsides are associated based on each lateral error to obtain each successfully associated roadside associated pair, it also includes: determining whether there is a third fused roadside that has not been successfully associated among each third fused roadside; in response to the existence of a third fused roadside that has not been successfully associated, the third fused roadside that has not been successfully associated is used as the fourth fused roadside, In response to the presence of a first roadside line that is not successfully associated with the fourth fused roadside line in the first preset number of associations in the future, the fourth fused roadside line is deleted; and / or, and / or, it is determined whether there is a first roadside line that is not successfully associated among the first roadside lines; in response to the presence of a first roadside line that is not successfully associated, the unsuccessfully associated first roadside line is used as a candidate tracking roadside line; in response to the presence of a first roadside line associated with the candidate tracking roadside line in the second preset number of associations in the future, the candidate tracking roadside line is determined to be a tracking roadside line; wherein, the first fused roadside line and the third fused roadside line both belong to tracking roadside lines.

[0240] In some embodiments, searching for detection roadside points of the same category as the target detection roadside points from a grid search range includes: determining the number of detection roadside points in the grid search range; in response to the number being greater than a preset number, treating each detection roadside point in the grid search range as a detection roadside point of the same category as the target detection roadside point; in response to the number being less than or equal to the preset number, treating each detection roadside point in the grid search range as a noise point; wherein, the grid search range is any one of a first grid search range and a second grid search range, in response to the grid search range being the first grid search range, the target detection roadside point represents the first roadside point; in response to the grid search range being the second grid search range, the target detection roadside point represents the second roadside point.

[0241] The above scheme performs rasterization processing on multiple detected curb points of the road at the current moment to obtain a corresponding grid map. The detected curb points can be placed in the grid map, and then clustered using the position of the grid where each detected curb point is located in the grid map to obtain multiple target curb point sets for determining several roadside lines at the current moment. Since the present application can directly use the position of the grid for clustering without considering the position of each detected curb point in the grid, compared to the method that needs to use the position of each point for clustering, the method of clustering using the grid position in the present application is conducive to improving the efficiency of clustering to obtain multiple target curb point sets, that is, it can improve the efficiency of determining the roadside line, and thus is conducive to improving the subsequent curb tracking efficiency.

[0242] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0243] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0244] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0245] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0246] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0247] If the integrated unit is implemented in the form of a software functional unit and 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 the present application is essentially 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, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0248] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A roadside tracking method, characterized in that: The method comprises: Perform rasterization processing on multiple detected roadside points of the current road to obtain a corresponding raster map; Clustering is performed using the positions of the grids where the detected roadside points are located in the grid map to obtain a plurality of target roadside point sets; the plurality of target roadside point sets are used to determine a plurality of first roadside lines at the current moment; Based on the first fused roadside at the historical moment and a number of first roadsides at the current moment, the second fused roadside at the current moment is obtained.

2. The method according to claim 1, characterized in that The clustering is performed using the positions of the grids where the detected roadside points are located in the grid map to obtain multiple target roadside point sets, including: Selecting an unsearched roadside point from a plurality of detected roadside points as a first roadside point; Searching for detected roadside points of the same category as the first roadside point from a first grid search range to obtain an initial roadside point set corresponding to the first roadside point; the first grid search range is determined based on a first position of a first grid where the first roadside point is located in the grid map, and is used to represent a grid range to be searched in the grid map; Based on the initial roadside point set corresponding to the first roadside point, determine a target roadside point set corresponding to the first roadside point; the target roadside point set includes all detected roadside points of the same category as the first roadside point among all the detected roadside points; Repeat the above steps until all detected roadside points in the grid map have been searched, and obtain a target roadside point set corresponding to each of the first roadside points.

3. The method according to claim 2, characterized in that The determining, based on the initial roadside point set corresponding to the first roadside point, a target roadside point set corresponding to the first roadside point comprises: Each detected roadside point in the initial roadside point set except the first roadside point is respectively used as a second roadside point; For each of the second roadside points, searching for a detection roadside point of the same category as the second roadside point from a second grid search range, and adding the searched detection roadside point to the initial roadside point set; the second grid search range is determined based on a second position of the second grid where the second roadside point is located in the grid map; Each searched detection roadside point is respectively used as a new second roadside point, and the step of searching for detection roadside points of the same category as the second roadside point from the second grid search range for each second roadside point is returned, and the searched detection roadside points are added to the initial roadside point set, until there are no detection roadside points of the same category as the second roadside points in the second grid search range, or the number of detection roadside points of the same category as the second roadside points searched is less than a set threshold, and the last determined initial roadside point set is used as the target roadside point set corresponding to the first roadside point.

4. The method according to claim 2, characterized in that: The grid map includes a plurality of grids distributed in rows and columns, and the position of the grid where the detection roadside point is located in the grid map includes the row and column value of the grid in the grid map; The detected roadside points of the same category as the first roadside point among all the detected roadside points include: the second roadside points located in the first grid search range corresponding to the first roadside point, and the detected roadside points located in the second grid search range corresponding to the second roadside points; The first grid search range is determined by using the first row and column value of the first grid where the first roadside point is located in the grid map and a preset search range value; The second grid search range is determined based on a second row and column value of the second grid where the second roadside point is located in the grid map and the preset search range value.

5. The method according to claim 4, characterized in that The preset search range value is smaller than the width of the road; The step of determining the grid search range includes: Determining the number of grid searches using the preset search range value and the resolution of the grid map; Determining the grid search range based on the target row and column values ​​of the target grid in the grid map and the grid search quantity; Among them, the grid search range is any one of the first grid search range and the second grid search range, in response to the grid search range being the first grid search range, the target grid represents the first grid where the first roadside point is located; in response to the grid search range being the second grid search range, the target grid represents the second grid where the second roadside point is located.

6. The method according to claim 5, characterized in that The target row and column values ​​include a target row value and a target column value of the target grid in the grid map; The step of determining the grid search range based on the target row and column value of the target grid in the grid map and the grid search quantity includes: In response to a first difference between the target row value and the grid search quantity being not less than a minimum row value, taking a first numerical value corresponding to the first difference as the minimum grid search row value, otherwise, taking the minimum row value as the minimum grid search row value; In response to a first summation result of the target row value and the grid search quantity being not greater than a maximum row value, taking a second value corresponding to the first summation result as a maximum grid search row value, otherwise, taking the maximum row value as the maximum grid search row value; In response to a second difference between the target column value and the grid search quantity being not less than a minimum column value, taking a third value corresponding to the second difference as the minimum grid search column value, otherwise, taking the minimum column value as the minimum grid search column value; In response to a second summation result of the target column value and the grid search quantity being not greater than a maximum column value, taking a fourth value corresponding to the second summation result as a maximum grid search column value, otherwise, taking the maximum column value as the maximum grid search column value; The minimum grid search row value, the maximum grid search row value, the minimum grid search column value and the maximum grid search column value are respectively used as search boundary values ​​to construct the grid search range; wherein the minimum row value, the maximum row value, the minimum column value and the maximum column value are respectively the map boundary values ​​of the grid map.

7. The method according to claim 1, characterized in that Determine several first routes at the current moment, including: Acquire second roadside lines determined based on each of the target roadside point sets; Determining target roadside line pairs satisfying a splicing condition from each of the second roadside lines; Based on the two second roadside lines in each of the target roadside line pairs, splicing is performed to obtain a corresponding spliced ​​roadside line; Each of the spliced ​​roadside lines is used as a first roadside line, and each remaining second roadside line that does not meet the splicing condition among the second roadside lines is used as a first roadside line, so as to obtain a plurality of the first roadside lines.

8. The method according to claim 7, characterized in that The step of determining target roadside line pairs satisfying a splicing condition from the second roadside lines includes: Selecting candidate roadside line pairs to be matched from each of the second roadside lines; For each of the candidate roadside line pairs, determining whether the two second roadside lines in the candidate roadside line pair meet a first splicing condition based on a first position difference between the two second roadside lines; In response to satisfying the first splicing condition, determining that the two second road lines satisfy the splicing condition; or, In response to satisfying the first splicing condition, determining whether the two second roadside lines satisfy a second splicing condition based on the obstacle detection result at the current moment; in response to the two second roadside lines satisfying the second splicing condition, determining that the two second roadside lines satisfy the splicing condition; The two second roadside lines that meet the splicing condition are used as two second roadside lines in a target roadside line pair.

9. The method according to claim 8, characterized in that The first position difference includes at least one of a lateral position difference and an angle difference; the first splicing condition includes that the lateral position difference is smaller than a lateral threshold, and / or the angle difference is smaller than an angle threshold.

10. The method according to claim 9, characterized in that Determining the lateral difference includes: Using one of the two second roadside lines as a first target roadside line, using the other as a second target roadside line, and selecting a target point from the first target roadside line; Acquire an associated target point of the target point, where the associated target point is obtained by extending the second target roadside line, and a longitudinal position difference between the target point and the associated target point is less than a preset difference threshold; Taking the lateral position difference between the target point and the associated target point as the lateral position difference between the two second road lines; Determining the angle difference includes: Obtaining a slope representation value of each of the two second roadside lines; The difference between the slope characterization values ​​respectively corresponding to the two second roadside lines is used as the angle difference between the two second roadside lines.

11. The method according to claim 10, characterized in that The slope characterization value corresponding to the first target roadside line is the first slope value of the curve equation corresponding to the first target roadside line at the target point, and the slope characterization value corresponding to the second target roadside line is the second slope value of the curve equation corresponding to the second target roadside line at the associated target point; The curve equation corresponding to each of the second roadside lines is obtained by performing curve fitting on the corresponding target roadside point set; And / or, the target point is a detected roadside point on the first target roadside line close to the second target roadside line.

12. The method according to claim 8, characterized in that The obstacle detection result includes one of a first detection result in which no obstacle exists and a second detection result in which an obstacle exists; The determining whether the two second roads meet the second splicing condition based on the obstacle detection result at the current moment includes: In response to the detection result being the first detection result, determining that the two second road edges do not satisfy the second splicing condition; In response to the detection result being the second detection result, determining a second position difference between the shielding area corresponding to the obstacle and the two second roadside lines; Based on the second position difference, it is determined whether the two second road lines meet a second splicing condition.

13. The method according to claim 12, characterized in that The second detection result includes a first position of an obstacle detection frame, and the occlusion area is determined by using the first position of the obstacle detection frame; The determining of the second position difference between the shielding area corresponding to the obstacle and the two second road lines includes: Determine first intersection points of the two second roadside lines or corresponding extension lines with the shielding area respectively; taking the longitudinal position difference between the target roadside point on each of the second roadside lines and the first intersection point of the corresponding second roadside line and the shielding area as the second position difference; The second splicing condition includes: each of the longitudinal position differences is not greater than a longitudinal difference threshold.

14. The method according to claim 13, characterized in that The second roadside line includes a starting roadside point and an ending roadside point, and the starting roadside point and the ending roadside point are determined based on the position sorting of each detected roadside point in the target roadside point set corresponding to the second roadside line; The target curb point on the front curb line of the two second curb lines is the starting curb point on the front curb line, and the target curb point on the rear curb line is the ending curb point on the rear curb line.

15. The method according to claim 13, characterized in that The occlusion area corresponding to the obstacle is determined by using the second positions corresponding to the four corner points of the obstacle detection frame, and the second position corresponding to each corner point is determined based on the first position of the obstacle detection frame; Determining the blocked area includes: In response to the obstacle being located on one side of the vehicle's forward direction, based on the second positions of the preset position points and the corner points on the vehicle at the current moment, constructing connecting lines between the preset position points and the corner points; Obtaining the angles between each of the connecting lines and the vehicle's forward direction, and taking the connecting line corresponding to the smallest of the angles as the first connecting line, and taking the connecting line corresponding to the largest of the angles as the second connecting line; Using the area between the first connecting line and the second connecting line as the shielding area; The first intersection points of the two second roadside lines or the corresponding extension lines with the shielding area include: the second intersection points of the two second roadside lines or the corresponding extension lines with the first connecting line, and the third intersection points of the two second roadside lines or the corresponding extension lines with the second connecting line.

16. The method according to claim 7, characterized in that The obtaining of the second roadside lines determined based on the target roadside point sets respectively comprises: Performing first curve fitting using each of the target roadside point sets respectively to obtain a second roadside line corresponding to each of the target roadside point sets; And / or, the step of splicing two second roadside lines in each of the target roadside line pairs to obtain a corresponding spliced ​​roadside line includes: For each of the target roadside line pairs, merging the target roadside point sets corresponding to each second roadside line in the target roadside line pair to obtain a merged roadside point set; A second curve fitting is performed on the combined roadside point set to obtain a spliced ​​roadside line.

17. The method according to claim 1, characterized in that The number of the first fusion roadside is at least one; The obtaining of the second fused roadside at the current moment based on the first fused roadside at the historical moment and the plurality of first roadside at the current moment comprises: Based on the driving data of the vehicle at the historical moment, predict the third fused roadside at the current moment corresponding to each first fused roadside at the historical moment; Associating each of the third fused roadsides with a plurality of the first roadsides, and determining associated pairs of roadsides that are successfully associated; For each of the roadside association pairs, the third fused roadside line and the first roadside line in the roadside association pair are fused to obtain the second fused roadside line at the current moment.

18. The method according to claim 17, characterized in that The step of associating each of the third fused roadside lines with a plurality of the first roadside lines to determine the roadside line association pairs that are successfully associated includes: Obtaining lateral errors between each of the third fused roadside lines and each of the first roadside lines; Roadside line association is performed based on each of the lateral errors to obtain successfully associated roadside line association pairs.

19. The method according to claim 18, characterized in that The step of associating the roadside lines based on the lateral errors to obtain the successfully associated roadside line association pairs comprises: Based on the lateral errors between each of the third fused roadside lines and each of the first roadside lines, constructing a cost matrix for characterizing each of the lateral errors; Performing roadside line pairing based on the cost matrix, selecting a roadside line target matching combination from a plurality of roadside line candidate matching combinations; each of the roadside line candidate matching combinations is a pairing combination including each of the third fused roadside lines and each of the first roadside lines; From the roadside line matching pairs in the roadside line target matching combination, each roadside line association pair that meets the preset requirements is selected; the lateral error between the third fused roadside line in each roadside line association pair and the corresponding first roadside line is less than a preset threshold.

20. The method according to claim 19, characterized in that The sum of first errors of all roadside line matching pairs in the roadside line target matching combination is less than the sum of second errors of all roadside line matching pairs in each of the roadside line candidate matching combinations; the sum of errors is the sum of lateral errors corresponding to all roadside line matching pairs; After the roadside lines are associated based on the lateral errors to obtain the successfully associated roadside line association pairs, the method further includes: Determine whether there is a third fusion route that has not been successfully associated among the third fusion routes; in response to the existence of a third fusion route that has not been successfully associated, use the third fusion route that has not been successfully associated as a fourth fusion route, and in response to the absence of a first route that has been successfully associated with the fourth fusion route in the first preset number of associations in the future, delete the fourth fusion route; and / or, and / or, determining whether there is a first roadside line that has not been successfully associated among the first roadside lines; in response to the existence of a first roadside line that has not been successfully associated, taking the first roadside line that has not been successfully associated as a candidate tracking roadside line, and in response to the existence of a first roadside line associated with the candidate tracking roadside line in a future second preset number of associations, determining the candidate tracking roadside line as a tracking roadside line; Among them, the first fused roadside line and the third fused roadside line are both tracking roadside lines.

21. The method according to claim 3, characterized in that Search for detection roadside points of the same category as the target detection roadside points in the grid search range, including: Determine the number of detected roadside points in the grid search range; In response to the number being greater than a preset number, each detected roadside point in the grid search range is regarded as a detected roadside point of the same category as the target detected roadside point; in response to the number being less than or equal to the preset number, each detected roadside point in the grid search range is regarded as a noise point; Among them, the grid search range is any one of the first grid search range and the second grid search range, in response to the grid search range being the first grid search range, the target detection roadside point represents a first roadside point; in response to the grid search range being the second grid search range, the target detection roadside point represents a second roadside point.

22. An electronic device, characterized in that: comprising a memory and a processor coupled to each other, The memory stores program instructions; The processor is used to execute program instructions stored in the memory to implement the method according to any one of claims 1 to 21.

23. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions that can be run by a processor, and the program instructions can be executed by the processor to implement the method according to any one of claims 1 to 21.

Images