Lane line generation method, device, electronic device and storage medium

By identifying and combining lane line point cloud data in gridded point cloud data, the target lane line is generated, and the problems of large error, low efficiency and high cost of lane line generation in the existing technology are solved, and the safety of autonomous driving cars is improved Sexual and economic benefits.

CN114241083BActive Publication Date: 2024-11-15BEIJING SAIMO TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111506915.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-10
Publication Date
2024-11-15
Estimated Expiration
2041-12-10

AI Technical Summary

Technical Problem

In the prior art, the lane line generation method has large errors, low efficiency and high economic costs, resulting in safety hazards when driving on the road.

Method used

By identifying the lane line point cloud data of each grid in the gridded point cloud data, candidate lane lines are generated, and candidate lane lines are merged to automatically generate target lane lines, reducing the repetition rate, improving the drawing accuracy, and reducing economics cost.

Benefits of technology

It improves the accuracy of drawing lane lines, reduces the safety risks of autonomous vehicles during driving on the road, and reduces the economic cost of drawing lane lines.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114241083B_ABST
    Figure CN114241083B_ABST
Patent Text Reader

Abstract

The present application provides a lane line generation method, device, electronic device and storage medium. The lane line generation method includes: generating gridded point cloud data; for each grid, identifying the point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generating candidate lane lines corresponding to each grid according to the point cloud data belonging to the lane line; based on the candidate lane lines, determining multiple first intersections; generating target lane lines according to the distances between the first intersections. The present application generates candidate lane lines by identifying the point cloud data of the lane line of each grid in the gridded point cloud data, and merging the candidate lane lines to generate the target lane line. On the basis of reducing the repetition rate of the target lane line, the lane line drawing accuracy is improved, and the economic cost of drawing the lane line is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a lane line generation method, device, electronic device and storage medium. Background Art

[0002] With the progress of society and the development of science and technology, self-driving cars, as a product of artificial intelligence development, have attracted much attention from artificial intelligence developers. The self-driving simulation system installed on self-driving cars needs to use lane lines to provide road protection and safety for automatic simulation driving.

[0003] Lane lines in the prior art are obtained by obtaining the coordinate data of all facets contained in dynamic objects in a virtual scene, filtering out the lane line data therefrom, and drawing the lane lines. The lane line data obtained directly from the coordinate data of all facets through simple screening will have certain errors, resulting in certain safety hazards for autonomous vehicles while driving on the road. In addition, the efficiency of obtaining lane line data by this method is low and the economic cost is high. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a lane line generation method, device, electronic device and storage medium, which generates candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merges the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of autonomous driving vehicles during driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0005] The present application provides a lane line generation method, the lane line generation method comprising:

[0006] During the driving process of the target vehicle, grids are added to the multi-frame point cloud data of the surrounding environment of the target vehicle to generate gridded point cloud data;

[0007] For each grid, identify the point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generate a candidate lane line corresponding to each grid based on the point cloud data belonging to the lane line;

[0008] For every two adjacent grids, determine a candidate lane line that intersects with a common grid line between the two adjacent grids, and determine a plurality of first intersection points;

[0009] According to the target distance between the first intersection points in every two adjacent grids, the two candidate lane lines corresponding to the two first intersection points are merged to generate the target lane line.

[0010] Furthermore, generating a candidate lane line corresponding to each grid according to the point cloud data belonging to the lane line includes:

[0011] Based on the point cloud density in the point cloud data belonging to the lane line, the point cloud data belonging to the lane line is clustered to generate candidate lane line point cloud data corresponding to each grid;

[0012] An initial fit is performed on the candidate lane line point cloud data corresponding to each grid to generate a candidate lane line corresponding to each grid.

[0013] Furthermore, for every two adjacent grids, determining a candidate lane line intersecting with a common grid line between the two adjacent grids, and determining a plurality of first intersection points, includes:

[0014] For every two adjacent grids, determine whether the candidate lane lines in the two adjacent grids intersect with the common grid line of the two adjacent grids;

[0015] If they intersect, then determine the candidate lane line that intersects with the common grid line between the two adjacent grids;

[0016] The intersection points between the plurality of intersecting candidate lane lines and the common grid line are determined as a plurality of first intersection points.

[0017] Furthermore, the step of merging two candidate lane lines corresponding to two first intersection points according to the target distance between the first intersection points in each two adjacent grids to generate the target lane line includes:

[0018] According to the target distance between each first intersection point in each two adjacent grids, determining whether the target distance is less than or equal to a first preset distance;

[0019] After determining that the distance is less than or equal to the first preset distance, based on each of the two first intersections, two candidate lane lines corresponding to each of the two first intersections are merged to generate a target lane line.

[0020] Further, after determining that the distance is less than or equal to the first preset distance, based on each of the two first intersections, merging two candidate lane lines corresponding to each of the two first intersections to generate a target lane line includes:

[0021] After determining that the distance is less than or equal to the first preset distance, determining an area formed by all candidate lane line point cloud data within a second preset distance from each of the first intersections as a target area;

[0022] Determine an intersection point between the two candidate lane lines corresponding to the two first intersection points and the target area as a second intersection point;

[0023] Target fitting is performed on the two second intersection points to generate a target lane line.

[0024] Furthermore, for each grid, identifying point cloud data belonging to the lane line from the point cloud data corresponding to each grid includes:

[0025] For each grid, identify the point cloud data belonging to the ground from the point cloud data corresponding to each grid;

[0026] The reflection intensity of each point in the point cloud data belonging to the ground in each grid is determined, and the ground point cloud data whose reflection intensity is greater than the preset reflection intensity is determined to be the point cloud data belonging to the lane line.

[0027] Furthermore, the multi-frame point cloud data includes global coordinates of the multi-frame point cloud, and the global coordinates of the multi-frame point cloud are generated by:

[0028] Acquire each frame of point cloud of the surrounding environment of the target vehicle, and generate initial local coordinate information of each frame of the point cloud based on the positioning information of each frame of the point cloud;

[0029] Based on the position and posture of the target vehicle during driving, motion de-distortion processing is performed on the initial local coordinate information of each frame of the point cloud to determine the instantaneous position information of each frame of the point cloud;

[0030] Determine the target local coordinate information of the point cloud of each frame according to the initial local coordinate information of the point cloud of each frame and the instantaneous position information of the point cloud of each frame;

[0031] The global coordinate information of the multi-frame point cloud is generated according to the target local coordinate information of each frame point cloud and the world coordinate system.

[0032] The embodiment of the present application further provides a lane line generating device, the lane line generating device comprising:

[0033] A generation module, used for adding grids to the multi-frame point cloud data of the surrounding environment of the target vehicle collected during the driving process of the target vehicle to generate gridded point cloud data;

[0034] A recognition module, for each grid, identifying point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generating a candidate lane line corresponding to each grid according to the point cloud data belonging to the lane line;

[0035] A determination module, configured to determine, for each two adjacent grids, a candidate lane line intersecting with a common grid line between the two adjacent grids, and determine a plurality of first intersection points;

[0036] The merging module is used to merge two candidate lane lines corresponding to each two first intersection points according to the target distance between each first intersection point in each two adjacent grids to generate a target lane line.

[0037] An embodiment of the present application also provides an electronic device, including: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the lane line generation method as described above are performed.

[0038] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the lane line generation method as described above are executed.

[0039] Compared with the prior art, the lane line generation method, device, electronic device and storage medium provided in the embodiments of the present application generate candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merge the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of autonomous driving vehicles during driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0040] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A flow chart of a lane line generation method provided in an embodiment of the present application is shown;

[0043] Figure 2 A flow chart of another lane line generation method provided in an embodiment of the present application is shown;

[0044] Figure 3 A schematic diagram of the structure of a lane line generating device provided in an embodiment of the present application is shown;

[0045] Figure 4A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown.

[0046] In the figure, 300 - lane line generating device; 310 - generating module; 320 - identifying module; 330 - determining module; 340 - merging module; 400 - electronic device; 410 - processor; 420 - memory; 430 - bus. DETAILED DESCRIPTION

[0047] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0048] First, research has found that lane lines in the prior art are obtained by obtaining the coordinate data of all facets contained in dynamic objects in the virtual scene, filtering out the lane line data, and drawing the lane lines. The lane line data directly obtained from the coordinate data of all facets through simple screening will have certain errors, resulting in certain safety hazards for autonomous vehicles while driving on the road. In addition, the efficiency of obtaining lane line data by this method is low and the economic cost is high.

[0049] Based on this, the embodiments of the present application provide a lane line generation method, device, electronic device and storage medium, which generates candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merges the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of autonomous driving vehicles during driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0050] See also Figure 1 , Figure 1 This is a flow chart of a lane line generation method provided in an embodiment of the present application. Figure 1 As shown in , the lane line generation method provided in the embodiment of the present application includes:

[0051] S101. When a target vehicle is driving, grids are added to multi-frame point cloud data of the target vehicle's surrounding environment to generate gridded point cloud data.

[0052] In this step, in the simulation test of the target vehicle's autonomous driving, first, it is necessary to collect multi-frame point cloud data about the surrounding environment of the target vehicle during its driving process, and add grids to the multi-frame point cloud data after collection to generate gridded point cloud data. The gridded point cloud data is more conducive to the management of lane lines.

[0053] Here, no matter what kind of data, after gridding processing is performed, more precise operations can be performed on each grid, avoiding the tediousness of directly processing a large amount of point cloud data, and compared with the data that has not been gridded, the gridded data can be used to realize precise operations on the data. Among them, the gridded point cloud data in this embodiment is a grid of 100*100 specifications, and the remaining point cloud data less than 100*100 can be determined to be added as a grid.

[0054] Among them, a high-line mechanical laser radar is used to collect multi-frame point cloud data. There is no limit on the number of lines of the high-line mechanical laser radar. Here, the high-line mechanical laser radar can obtain a three-dimensional model of the target vehicle's surrounding environment through laser scanning, and compare the environment when the previous frame of data was collected with the environment when the next frame of data was collected. Through changes in the environment, all static or dynamic point cloud data such as surrounding vehicles, pedestrians, and lane lines can be easily detected.

[0055] In this way, point cloud data refers to a collection of vectors in a three-dimensional coordinate system. Point cloud data is obtained by using the high-line mechanical laser radar provided in this embodiment to scan the surface of an object and record all partial data of the object in the form of points. Each point data contains three-dimensional coordinates, and some may contain color (RGB) information or reflection intensity information.

[0056] Furthermore, the multi-frame point cloud data includes global coordinates of the multi-frame point cloud, and the global coordinates of the multi-frame point cloud are generated by:

[0057] Each frame of point cloud of the target vehicle's surrounding environment is acquired, and based on the positioning information of each frame of point cloud, initial local coordinate information of each frame of the point cloud is generated.

[0058] Here, all the patch data collected by the high-line mechanical laser radar installed on the top of the target vehicle during one rotation is defined as each frame of point cloud. Based on the positioning information of each frame of point cloud collected by the combined navigation system, the initial local coordinate information of each frame of point cloud is generated, wherein the origin in the initial local coordinate information of each frame of point cloud is the current position coordinate of the target vehicle in each frame.

[0059] In this way, the combined navigation system provides high-precision positioning information for each frame of point cloud. The combined navigation system provided in this embodiment includes but is not limited to the use of a global navigation satellite system (GNSS), an inertial measurement unit (IMU) and a real-time dynamic positioning sensor (RTK). Among them, the global navigation satellite system is an air-based radio navigation and positioning system that can provide users with all-weather three-dimensional coordinates, speed and time information at any location on the surface of the earth or in near-Earth space; inertial sensors are mainly used to detect and measure acceleration and rotational motion; real-time dynamic positioning sensors are a measurement method that can obtain centimeter-level positioning accuracy in real time in the field. Its appearance has greatly improved outdoor work efficiency.

[0060] Based on the position and posture of the target vehicle during driving, motion de-distortion processing is performed on the initial local coordinate information of each frame of the point cloud to determine the instantaneous position information of each frame of the point cloud.

[0061] Here, the combined navigation system is used to obtain the position and posture of the target vehicle during driving. Since the target vehicle will collect each frame of point cloud data in real time through the high-line mechanical laser radar during driving, and each frame of point cloud data is obtained while the target vehicle keeps driving, that is to say, the initial local coordinate information of each frame of the point cloud obtained here is not obtained at the instantaneous moment, and when the frequency of the high-line mechanical laser radar scanning rotation is relatively low, the motion error of the laser frame caused by the movement of the target vehicle cannot be ignored. Therefore, there is a motion distortion error, which causes the distortion of the initial local coordinate information. Therefore, it is necessary to perform motion de-distortion processing on the initial local coordinate information of each frame of the point cloud based on the position and posture of the target vehicle during driving. The laser radar coordinates corresponding to each laser point in each frame of point cloud data collected by the high-line mechanical laser radar are converted to the relative instantaneous position, and the instantaneous position information of each frame of point cloud is determined to achieve the purpose of motion distortion removal.

[0062] Among them, the embodiments of motion distortion generation can be obtained through the following embodiments provided in this application:

[0063] When the scanning frequency of the high-line mechanical laser radar installed on the top of the target vehicle is 5Hz, the end time difference of each frame of point cloud data is 200ms. If the robot walks in the preset direction at a speed of 0.5m / s and scans the road in front, then after 200ms, the measured distance of the tail and the measured distance of the head in the preset direction will differ by 10cm.

[0064] According to the initial local coordinate information of the point cloud in each frame and the instantaneous position information of the point cloud in each frame, the target local coordinate information of the point cloud in each frame is determined.

[0065] The global coordinate information of the multi-frame point cloud is generated according to the target local coordinate information of each frame point cloud and the world coordinate system.

[0066] Here, the target local coordinate information of each frame point cloud is the local coordinate in the local coordinate system with the current position of the target vehicle as the origin of the coordinate system. In order to facilitate the use and testing of the generated target lane line, the local coordinate information needs to be converted into global coordinates in the global coordinate system, and the global coordinate information of multiple frames of point clouds is generated.

[0067] Among them, the method of converting local coordinate information into global coordinates under the global coordinate system is to put the target local coordinate information of each frame point cloud into the world coordinate system with the starting point of the target vehicle as the origin, generate the global coordinate information of multiple frames of point clouds, and mark the global coordinate information of each grid in the lower left corner of each grid to facilitate search and query in the subsequent tracing process.

[0068] S102. For each grid, identify point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generate candidate lane lines corresponding to each grid based on the point cloud data belonging to the lane line.

[0069] In this step, the point cloud data corresponding to each grid includes all the patch point cloud data of dynamic objects and static objects. It is necessary to identify the patch point cloud data from a large number of patch point cloud data, and then generate the candidate lane line corresponding to each grid based on the point cloud data of the lane line.

[0070] Furthermore, for each grid, identifying point cloud data belonging to the lane line from the point cloud data corresponding to each grid includes:

[0071] For each grid, point cloud data belonging to the ground is identified from the point cloud data corresponding to each grid.

[0072] Here, the point cloud data corresponding to the largest facet is obtained from the point cloud data corresponding to each grid through the Random Sample Consensus (RANSAC) algorithm, and the above point cloud data is determined as point cloud data belonging to the ground. Here, the Random Sample Consensus algorithm is an algorithm that calculates the mathematical model parameters of the data based on a set of sample data sets containing abnormal data to obtain valid sample data.

[0073] The reflection intensity of each point in the point cloud data belonging to the ground in each grid is determined, and the ground point cloud data whose reflection intensity is greater than the preset reflection intensity is determined to be the point cloud data belonging to the lane line.

[0074] Among them, the point cloud data belonging to the ground includes not only the point cloud data belonging to the lane lines, but also other types of ground point cloud data. The lane lines on the ground are mostly marked lines with high color brightness and reflection intensity, such as yellow or white marked lines, while the ground at other positions outside the lane lines is mostly with low color brightness and reflection intensity, for example, dark or dark-colored ground. Therefore, the reflection intensity of each point in the point cloud data belonging to the ground in each grid can be obtained, and the ground point cloud data with a reflection intensity greater than the preset reflection intensity can be determined as the point cloud data belonging to the lane lines.

[0075] In this way, the minimum value of the reflection intensity of the lane line point cloud data can be obtained by testing the sample lane line, and the minimum value can be set as the preset reflection intensity.

[0076] S103 . For every two adjacent grids, determine a candidate lane line that intersects with a common grid line between the two adjacent grids, and determine a plurality of first intersection points.

[0077] In this step, two adjacent grids are selected from the multiple added grids, and it is determined whether there is a candidate lane line in each of the two adjacent grids that intersects with the grid. After determining that there is a candidate lane line in each of the two adjacent grids that intersects with the grid, it is determined whether each intersection point of each candidate lane line with the grid is located on a common grid line between the two adjacent grids. If each intersection point with the grid is located on the common grid line between the two adjacent grids, the multiple intersection points on the common grid line are determined to be multiple first intersection points.

[0078] Here, the common grid line between two adjacent grids is a grid line shared by the two adjacent grids.

[0079] Furthermore, for every two adjacent grids, determining a candidate lane line intersecting with a common grid line between the two adjacent grids, and determining a plurality of first intersection points, includes:

[0080] For every two adjacent grids, determine whether the candidate lane lines in the two adjacent grids intersect with a common grid line of the two adjacent grids.

[0081] Here, the candidate lane lines in the two adjacent grids may intersect with the common grid line of the two adjacent grids, or may not intersect with the common grid line of the two adjacent grids. Here, the non-intersecting candidate lane line may be a shorter target lane line.

[0082] If they intersect, a candidate lane line intersecting with the common grid line between the two adjacent grids is determined.

[0083] The intersection points between the plurality of intersecting candidate lane lines and the common grid line are determined as a plurality of first intersection points.

[0084] S104: According to the target distance between the first intersection points in every two adjacent grids, two candidate lane lines corresponding to the two first intersection points are merged to generate a target lane line.

[0085] In this step, the target distance between the first intersections in every two adjacent grids is calculated, and the target distance between the first intersections is judged. When the target distance between the first intersections meets the condition of the first preset distance, the two candidate lane lines corresponding to the two first intersections can be merged, and the two candidate lane lines can be treated as the same lane line to generate the target lane line, and the connection information in the adjacent grids is attached to indicate the position of the current lane line on the adjacent grids, thereby achieving deduplication of the candidate lane lines.

[0086] Compared with the prior art, the lane line generation method provided in the embodiment of the present application generates candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merges the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of autonomous driving vehicles during driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0087] See also Figure 2 , Figure 2 This is a flow chart of a lane line generation method provided by another embodiment of the present application. Figure 2 As shown in , the lane line generation method provided in the embodiment of the present application includes:

[0088] S201. During the driving process of a target vehicle, grids are added to the collected multi-frame point cloud data of the surrounding environment of the target vehicle to generate gridded point cloud data.

[0089] S202: clustering the point cloud data belonging to the lane lines based on the point cloud density in the point cloud data belonging to the lane lines, and generating candidate lane line point cloud data corresponding to each grid.

[0090] In this step, a density clustering algorithm is used to cluster the point cloud data belonging to the lane lines, and candidate lane line point cloud data corresponding to each grid is generated. The candidate lane line point cloud data whose number is less than the preset number of point clouds are eliminated, thereby removing noise, impurities and some other non-lane line point cloud data, and finally generating candidate lane line point cloud data corresponding to each grid.

[0091] Among them, the density clustering algorithm provided in the embodiments of the present application includes but is not limited to the use of a density clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN). Here, the density clustering algorithm clustering method is different from the partitioning and hierarchical clustering methods. It defines a cluster as the largest set of density-connected points, can divide areas with sufficiently high density into clusters, and can find clusters of arbitrary shapes in a noisy spatial database.

[0092] S203: Perform initial fitting on the candidate lane line point cloud data corresponding to each grid to generate a candidate lane line corresponding to each grid.

[0093] In this step, the candidate lane line point cloud data is initially fitted using the least squares method to fit a smooth cubic curve as the candidate lane line.

[0094] Among them, fitting is a way of substituting the existing lane line point cloud data into a mathematical expression through mathematical methods. The least squares method is a mathematical optimization technology. It finds the best function match for the lane line point cloud data by minimizing the sum of squares of errors. The least squares method can be used to easily obtain unknown lane line point cloud data and minimize the sum of squares of errors between the obtained lane line point cloud data and the actual data.

[0095] In this way, the function expression of the least squares method is:

[0096] f(x)=α1ψ1(x)+α2ψ2(x)+…+αmψm(x);

[0097] Among them, ψm is a set of linearly independent functions selected in advance, and αm is the coefficient to be determined.

[0098] S204: For every two adjacent grids, determine a candidate lane line that intersects with a common grid line between the two adjacent grids, and determine a plurality of first intersection points.

[0099] S205 . According to the target distance between the first intersection points in every two adjacent grids, two candidate lane lines corresponding to the two first intersection points are merged to generate a target lane line.

[0100] Furthermore, the step of merging two candidate lane lines corresponding to two first intersection points according to the target distance between the first intersection points in each two adjacent grids to generate the target lane line includes:

[0101] According to the target distance between the first intersection points in every two adjacent grids, it is determined whether the distance is less than or equal to the first preset distance.

[0102] After determining that the distance is less than or equal to the first preset distance, based on each of the two first intersections, two candidate lane lines corresponding to each of the two first intersections are merged to generate a target lane line.

[0103] Further, after determining that the distance is less than or equal to the first preset distance, based on each of the two first intersections, merging two candidate lane lines corresponding to each of the two first intersections to generate a target lane line includes:

[0104] After determining that the distance is less than or equal to the first preset distance, an area formed by the point cloud data of all candidate lane lines within a second preset distance from each of the first intersections is determined as a target area.

[0105] Here, an area formed by point cloud data of all candidate lane lines within a second preset distance from each of the first intersection points is determined as a target area, wherein the shape of the target area includes but is not limited to a rectangle, a square, a circle, and the like.

[0106] An intersection point between the two candidate lane lines corresponding to the two first intersection points and the target area is determined as a second intersection point.

[0107] Here, the intersection point between the two candidate lane lines and the target area is taken as the second intersection point, and the second intersection point is used to fit the cut-off part in the target generation curve, so that the cut-off part in the target generation curve makes the curve smooth.

[0108] Target fitting is performed on the two second intersection points to generate a target lane line.

[0109] Here, a spline point set is used to generate a flexible band for smoothing the target lane line, and the flexible band and the two second intersection points are merged to generate the target lane line.

[0110] Among them, Spline is a drawing tool for drawing curves. It is a flexible and long strip. When drawing, use a pressure iron to make the spline pass through the specified sample points, adjust the spline to make it have a satisfactory shape, and then draw the curve along the spline.

[0111] Among them, the description of S201 and S204 to S205 can refer to the description of S101 and S103 to S104, and can achieve the same technical effect, which is not repeated here.

[0112] Compared with the prior art, the lane line generation method provided in the embodiment of the present application generates candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merges the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of autonomous driving vehicles during driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0113] See also Figure 3 , Figure 3 This is a schematic diagram of the structure of a lane line generating device provided in an embodiment of the present application. Figure 3 As shown in , the lane line generating device 300 includes:

[0114] A generating module 310 is used to add grids to the multi-frame point cloud data of the surrounding environment of the target vehicle collected during the driving process of the target vehicle to generate gridded point cloud data;

[0115] The identification module 320 is used to identify, for each grid, point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generate a candidate lane line corresponding to each grid according to the point cloud data belonging to the lane line;

[0116] A determination module 330 is used to determine, for each two adjacent grids, a candidate lane line that intersects with a common grid line between the two adjacent grids, and determine a plurality of first intersection points;

[0117] The merging module 340 is used to merge two candidate lane lines corresponding to each two first intersection points according to the distance between each first intersection point in each two adjacent grids to generate a target lane line.

[0118] Compared with the prior art, the lane line generating device 300 provided in the embodiment of the present application generates candidate lane lines by identifying the point cloud data of the lane lines of each grid in the gridded point cloud data, and merges the candidate lane lines to automatically generate target lane lines. On the basis of reducing the repetition rate of the target lane lines, the lane line drawing accuracy is improved, the safety hazards of the autonomous driving car in the process of driving on the road are reduced, and the economic cost of drawing lane lines is reduced.

[0119] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown in , the electronic device 400 includes a processor 410 , a memory 420 and a bus 430 .

[0120] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, the above-mentioned Figure 1 as well as Figure 2 The steps of the lane line generation method in the method embodiment shown, the specific implementation method can be found in the method embodiment, and will not be repeated here.

[0121] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 as well as Figure 2 The steps of the lane line generation method in the method embodiment shown, the specific implementation method can be found in the method embodiment, and will not be repeated here.

[0122] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0123] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, 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 communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0124] 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 solution of this embodiment.

[0125] 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.

[0126] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0127] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A lane line generation method, It is characterized in that The lane line generation method comprises: During the driving process of the target vehicle, grids are added to the multi-frame point cloud data of the surrounding environment of the target vehicle to generate gridded point cloud data; For each grid, identify the point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generate a candidate lane line corresponding to each grid based on the point cloud data belonging to the lane line; For every two adjacent grids, determine a candidate lane line that intersects with a common grid line between the two adjacent grids, and determine a plurality of first intersection points; According to the target distance between the first intersection points in each two adjacent grids, two candidate lane lines corresponding to the two first intersection points are merged to generate a target lane line; The step of determining, for each two adjacent grids, a candidate lane line intersecting with a common grid line between the two adjacent grids, and determining a plurality of first intersection points includes: For every two adjacent grids, determining whether the candidate lane lines in the two adjacent grids intersect with a common grid line of the two adjacent grids; If they intersect, then determine the candidate lane line that intersects with the common grid line between the two adjacent grids; Determining the intersections between a plurality of intersecting candidate lane lines and a common grid line as a plurality of first intersections; merging two candidate lane lines corresponding to each of the two first intersections according to the target distance between each of the first intersections in each of two adjacent grids to generate a target lane line, including: According to the target distance between each first intersection point in each two adjacent grids, determining whether the target distance is less than or equal to a first preset distance; After determining that the target distance is less than or equal to the first preset distance, based on each of the two first intersections, two candidate lane lines corresponding to each of the two first intersections are merged to generate a target lane line.

2. The lane line generation method according to claim 1, It is characterized in that The step of generating a candidate lane line corresponding to each grid according to the point cloud data belonging to the lane line includes: Based on the point cloud density in the point cloud data belonging to the lane line, the point cloud data belonging to the lane line is clustered to generate candidate lane line point cloud data corresponding to each grid; An initial fit is performed on the candidate lane line point cloud data corresponding to each grid to generate a candidate lane line corresponding to each grid.

3. The lane line generation method according to claim 2, It is characterized in that After determining that the target distance is less than or equal to the first preset distance, based on each of the two first intersections, merging two candidate lane lines corresponding to each of the two first intersections to generate a target lane line includes: After determining that the target distance is less than or equal to the first preset distance, determining an area formed by all candidate lane line point cloud data within a second preset distance from each of the first intersections as a target area; Determine an intersection point between the two candidate lane lines corresponding to the two first intersection points and the target area as a second intersection point; Target fitting is performed on the two second intersection points to generate a target lane line.

4. The lane line generation method according to claim 1, It is characterized in that For each grid, identifying point cloud data belonging to the lane line from the point cloud data corresponding to each grid includes: For each grid, identify the point cloud data belonging to the ground from the point cloud data corresponding to each grid; The reflection intensity of each point in the point cloud data belonging to the ground in each grid is determined, and the ground point cloud data whose reflection intensity is greater than the preset reflection intensity is determined to be the point cloud data belonging to the lane line.

5. The lane line generation method according to claim 1, It is characterized in that The multi-frame point cloud data includes the global coordinates of the multi-frame point cloud, and the global coordinates of the multi-frame point cloud are generated by the following method: Acquire each frame of point cloud of the surrounding environment of the target vehicle, and generate initial local coordinate information of each frame of the point cloud based on the positioning information of each frame of the point cloud; Based on the position and posture of the target vehicle during driving, motion de-distortion processing is performed on the initial local coordinate information of each frame of the point cloud to determine the instantaneous position information of each frame of the point cloud; Determine the target local coordinate information of the point cloud of each frame according to the initial local coordinate information of the point cloud of each frame and the instantaneous position information of the point cloud of each frame; The global coordinate information of the multi-frame point cloud is generated according to the target local coordinate information of each frame point cloud and the world coordinate system.

6. A lane line generating device, It is characterized in that The lane line generating device comprises: A generation module, used for adding grids to the multi-frame point cloud data of the surrounding environment of the target vehicle collected during the driving process of the target vehicle to generate gridded point cloud data; A recognition module, for each grid, identifying point cloud data belonging to the lane line from the point cloud data corresponding to each grid, and generating a candidate lane line corresponding to each grid according to the point cloud data belonging to the lane line; A determination module, configured to determine, for each two adjacent grids, a candidate lane line intersecting with a common grid line between the two adjacent grids, and determine a plurality of first intersection points; A merging module, configured to merge two candidate lane lines corresponding to each of the two first intersection points according to the target distance between each of the first intersection points in each of the two adjacent grids, to generate a target lane line; When the determination module is used to determine, for each two adjacent grids, a candidate lane line intersecting with a common grid line between the two adjacent grids and to determine a plurality of first intersection points, it is specifically used to: For each two adjacent grids, determine whether the candidate lane lines in the two adjacent grids intersect with the common grid line of the two adjacent grids; if so, determine the candidate lane lines that intersect with the common grid line between the two adjacent grids; determine the intersection points between multiple intersecting candidate lane lines and the common grid line as multiple first intersection points; The merging module is used to merge two candidate lane lines corresponding to each two first intersection points according to the target distance between each first intersection point in each two adjacent grids to generate a target lane line, specifically for: According to the target distance between the first intersection points in each two adjacent grids, determine whether the target distance is less than or equal to the first preset distance; after determining that the target distance is less than or equal to the first preset distance, based on each two first intersection points, merge the two candidate lane lines corresponding to each two first intersection points to generate a target lane line.

7. An electronic device, It is characterized in that include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via the bus, and the machine-readable instructions are executed by the processor to execute the steps of the lane line generation method as described in any one of claims 1 to 5 above.

8. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the lane line generation method as described in any one of claims 1 to 5 are executed.