CN116434189A

CN116434189A - Road edge detection method, device, equipment and storage medium

Info

Publication number: CN116434189A
Application number: CN202310343422.8A
Authority: CN
Inventors: 罗建国
Original assignee: Chongqing Changan Automobile Co Ltd
Current assignee: Chongqing Changan Automobile Co Ltd
Priority date: 2023-03-31
Filing date: 2023-03-31
Publication date: 2023-07-14

Abstract

The application relates to a road edge detection method, a road edge detection device, road edge detection equipment and a storage medium, and relates to the technical field of automobiles. The method comprises the following steps: constructing and obtaining a front view corresponding to the target area based on the plurality of point cloud data; determining a gradient value corresponding to each point cloud data based on the coordinate value corresponding to each point cloud data; constructing a gridding top view corresponding to the target area based on the plurality of point cloud data, and determining a grid characteristic value of each grid; and screening at least one point cloud data from the plurality of point cloud data based on the gradient value and the grid characteristic value corresponding to each point cloud data, performing curve fitting processing based on the at least one point cloud data, and determining the road edge position corresponding to the target area. Therefore, the road edge key points can be screened from the point cloud data through the front view and the top view, so that the accuracy and the robustness of road edge detection are improved, and the problem that the road edge key points are missed and misdetected when the point cloud data are screened only through the front view or the top view is solved.

Description

Road edge detection method, device, equipment and storage medium

Technical Field

The invention relates to the technical field of automobiles, in particular to a road edge detection method, a device, equipment and a storage medium.

Background

With the widespread application of the laser radar in the field of automatic driving perception, currently, when the perception algorithm based on the laser radar detects the road edge information, a large amount of point cloud data collected by the laser radar can be detected and filtered through front view gradient information or top view grid characteristics under a single view (front view or top view), so that a road edge candidate point is screened from the large amount of point cloud data, and the road edge candidate point is output through a curve fitting method, so that a road edge detection result is obtained.

However, when the front view gradient detection point cloud data is passed, the far-distance point cloud data and the blocked point cloud data cannot be detected, so that the missed detection of the road edge key point (namely, the non-ground point) is caused, meanwhile, the point cloud noise and the measurement error are sensitive, so that the false detection of the road edge key point is caused, and when the top view grid detection point cloud data is passed, the grid preset size is limited, the road edge key point which is the short road edge is screened out as the road surface point, so that the missed detection of the road edge key point is caused, and the accuracy of the road edge detection is lower, and the robustness is poor.

Disclosure of Invention

The invention aims to provide a method, a device, equipment and a storage medium for detecting a road edge, which are used for solving the technical problem of missing detection and false detection of a road edge key point when road edge information is detected only through a front view gradient or a top view grid. The technical scheme of the application is as follows:

According to a first aspect of the present application, there is provided a method for detecting a road edge, including: acquiring a plurality of point cloud data corresponding to a target area, and constructing a front view corresponding to the target area based on the plurality of point cloud data, wherein the plurality of point cloud data are acquired from the target area, and one point cloud data corresponds to one point in the front view; determining a gradient value corresponding to each point cloud data in the plurality of point cloud data based on the coordinate value corresponding to each point cloud data of the front view in the preset coordinate system; constructing a gridding top view corresponding to the target area based on the plurality of point cloud data, and determining a grid characteristic value of each grid included in the gridding top view, wherein the grid characteristic value comprises at least one of the following: the number of the point cloud data included in the grid and the polar coordinate difference value corresponding to the point cloud data included in the grid; and screening at least one point cloud data from the plurality of point cloud data based on the gradient value corresponding to each point cloud data in the plurality of point cloud data and the grid characteristic value of the grid corresponding to each point cloud data in the grid plan view, and performing curve fitting processing based on the screened at least one point cloud data to determine the road edge position corresponding to the target area.

According to the technical means, the effective point cloud data can be screened out through the gradient values corresponding to the point cloud data in the front view, the problem that the point cloud data formed by the short road edges or the fences cannot be completely displayed in the grid top view due to the limited grid size when the effective point cloud data are screened out in the top view, so that the short road edge key points or the fence key points are missed is solved, meanwhile, the effective point cloud data can be screened out through the grid characteristics of the point cloud data in the top view, and the problem that the gradient values corresponding to the point cloud data which are far away or are blocked in the front view cannot be acquired when the effective point cloud data are screened out in the front view is avoided, so that the problem of missing effective point cloud data is solved, and the accuracy and the robustness of road edge detection are improved.

In one possible implementation manner, based on a gradient value corresponding to each point cloud data in the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the grid plan view, at least one point cloud data is selected from the plurality of point cloud data, including: determining a plurality of first point cloud data with gradient values larger than a preset gradient threshold value corresponding to the point cloud data from the plurality of point cloud data; and determining at least one point cloud data of which the grid characteristic value of the grid corresponding to the point cloud data in the grid top view is larger than a preset characteristic value from the plurality of first point cloud data.

According to the technical means, the method and the device can screen the point cloud data based on the gradient value corresponding to the point cloud data in the front view and the grid characteristic value of the grid corresponding to the top view, and can solve the problem that the detection accuracy of the road edge detection is insufficient in a single view.

In one possible implementation, the polar coordinate difference value corresponding to the point cloud data included in the grid is: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate is further included in the method: determining a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view based on the coordinate value of each of the plurality of point cloud data in the rasterized top view; and determining a polar coordinate difference value corresponding to the point cloud data included in each grid on the ordinate based on the maximum ordinate value and the minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view.

According to the technical means, the polar coordinate difference value corresponding to the point cloud data on the ordinate can be determined based on the coordinate value of the point cloud data in the rasterized top view, so that the polar coordinate difference value is determined to be the grid characteristic value of the grid corresponding to the point cloud data in the rasterized top view, and further, effective point cloud data are screened out from the point cloud data based on the grid characteristic value corresponding to the point cloud data, and the accuracy of screening the road edge key points is improved.

In one possible implementation manner, performing curve fitting processing based on the screened at least one point cloud data to determine a path edge position corresponding to the target area includes: based on the grid characteristic values of grids corresponding to each point cloud data in the screened at least one point cloud data in the grid top view, clustering the grids in the grid top view to obtain multiple clusters; determining a cluster characteristic value corresponding to each type of cluster in the multi-type cluster, and screening at least one type of cluster meeting a preset condition from the multi-type cluster, wherein the cluster characteristic value comprises at least one of the following: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster, and the preset condition comprises at least one of the following: the clustering characteristic value is larger than a preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value; and performing curve fitting processing based on coordinate values corresponding to the point cloud data included in at least one type of cluster, and determining the road edge position corresponding to the target area.

According to the technical means, at least one type of cluster meeting the preset conditions can be screened from the multiple types of clusters based on the cluster characteristic values corresponding to the clusters, namely, the clusters formed by vehicles, pedestrians and the formed clusters are filtered from the multiple types of clusters, and the cluster formed by the road edges is reserved, so that non-road edge targets in the road can be rapidly and effectively eliminated, more accurate road edge targets are obtained, the problem of false road edge detection caused by the vehicles and the pedestrians in the road is solved, and the effective distance and the accuracy of the road edge detection can be improved by introducing the cluster characteristic values.

According to a second aspect provided by the present application, a road edge detection device is provided, which includes an acquisition module, a determination module and a processing module; the acquisition module is used for acquiring a plurality of point cloud data corresponding to the target area, wherein the plurality of point cloud data are acquired from the target area; the processing module is used for constructing and obtaining a front view corresponding to the target area based on a plurality of point cloud data, wherein one point cloud data corresponds to one point in the front view; the determining module is used for determining a gradient value corresponding to each point cloud data in the plurality of point cloud data based on the coordinate value corresponding to each point cloud data of the front view in the preset coordinate system; the processing module is also used for constructing a grid top view corresponding to the target area based on the plurality of point cloud data; the determining module is further configured to determine a grid characteristic value of each grid included in the rasterized top view, where the grid characteristic value includes at least one of: the number of the point cloud data included in the grid and the polar coordinate difference value corresponding to the point cloud data included in the grid; the processing module is further used for screening at least one point cloud data from the plurality of point cloud data based on the gradient value corresponding to each point cloud data in the plurality of point cloud data and the grid characteristic value of the grid corresponding to each point cloud data in the grid plan view, and performing curve fitting processing based on the screened at least one point cloud data; the determining module is further used for determining the route edge position corresponding to the target area.

In a possible implementation manner, the determining module is further configured to determine, from the plurality of point cloud data, a plurality of first point cloud data with gradient values corresponding to the point cloud data greater than a preset gradient threshold; the determining module is further configured to determine, from the plurality of first point cloud data, at least one point cloud data, where a grid characteristic value of a grid corresponding to the point cloud data in the grid top view is greater than a preset characteristic value.

In one possible implementation, the polar coordinate difference value corresponding to the point cloud data included in the grid is: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate; the determining module is further used for determining a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the grid top view based on the coordinate value of each point cloud data in the grid top view; the determining module is further configured to determine a polar coordinate difference value corresponding to the point cloud data included in each grid on the ordinate, based on a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view.

In a possible implementation manner, the processing module is further configured to perform clustering processing on the grids included in the top-view gridding plan view based on the grid characteristic values of the grids corresponding to each point cloud data in the screened at least one point cloud data in the top-view gridding plan view to obtain multiple types of cluster clusters; the determining module is further configured to determine a cluster feature value corresponding to each of the multiple types of clusters, where the cluster feature value includes at least one of: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster; the processing module is further used for screening at least one type of cluster meeting preset conditions from the multiple types of cluster, wherein the preset conditions comprise at least one of the following: the clustering characteristic value is larger than a preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value; the processing module is also used for performing curve fitting processing based on coordinate values corresponding to the point cloud data included in at least one type of cluster; the determining module is further used for determining the route edge position corresponding to the target area.

According to a third aspect provided by the present application, there is provided an electronic device comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute instructions to implement the method of the first aspect and any of its possible embodiments described above.

According to a fourth aspect provided herein, there is provided a computer readable storage medium, which when executed by a processor of an electronic device, enables the electronic device to perform the method of any one of the above-mentioned first aspects and any one of its possible embodiments.

According to a fifth aspect provided by the present application, there is provided a vehicle comprising: a road edge detection device for implementing the method of the first aspect and any possible implementation manner thereof.

According to a sixth aspect provided herein, there is provided a computer program product comprising computer instructions which, when run on an electronic device, cause the electronic device to perform the method of the first aspect and any one of its possible embodiments.

Therefore, the technical characteristics of the application have the following beneficial effects:

(1) Effective point cloud data can be screened out through corresponding gradient values of the point cloud data in the front view, the problem that when the effective point cloud data is screened out in the top view, due to the fact that the grid size is limited, the point cloud data formed by the short road edges or the fences cannot be all displayed in the grid top view, so that the problem that the short road edges or the fence key points are missed is solved, meanwhile, the effective point cloud data can be screened out through grid features of the point cloud data in the top view, and the problem that when the effective point cloud data is screened out in the front view, the gradient values corresponding to the far-distance or blocked point cloud data in the front view cannot be obtained, so that the problem that the effective point cloud data is missed is solved, and therefore the accuracy and the robustness of road edge detection are improved.

(2) The method can screen the point cloud data based on the gradient value corresponding to the point cloud data in the front view and the grid characteristic value of the grid corresponding to the top view, and can solve the problem of insufficient detection precision of the road edge detection in the single view.

(3) The polar coordinate difference value corresponding to the point cloud data on the ordinate can be determined based on the coordinate value of the point cloud data in the rasterized top view, so that the polar coordinate difference value is determined to be the grid characteristic value of the grid corresponding to the point cloud data in the rasterized top view, and further, effective point cloud data is screened out from the point cloud data based on the grid characteristic value corresponding to the point cloud data, and the accuracy of screening the road edge key points is improved.

(4) At least one type of cluster meeting preset conditions can be screened out from multiple types of cluster based on the cluster characteristic values corresponding to the cluster clusters, namely, the clusters formed by vehicles, pedestrians and the forming clusters are filtered out from the multiple types of cluster clusters, and the cluster clusters formed by the road edges are reserved, so that non-road edge targets in roads can be rapidly and effectively eliminated, more accurate road edge targets are obtained, the problem of false road edge detection caused by vehicles and pedestrians in the roads is solved, and the effective distance and the accuracy of road edge detection can be improved by introducing the cluster characteristic values.

It should be noted that, the technical effects caused by any implementation manner of the second aspect to the sixth aspect may refer to the technical effects caused by the corresponding implementation manner in the first aspect, which is not described herein.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application and do not constitute an undue limitation on the application.

FIG. 1 is a schematic diagram of a path edge detection system according to an exemplary embodiment;

FIG. 2 is a flow chart illustrating a method of detecting a road edge according to an example embodiment;

FIG. 3 is a flowchart illustrating yet another method of edge detection, according to an example embodiment;

FIG. 4 is a flowchart illustrating yet another method of edge detection, according to an example embodiment;

FIG. 5 is a flowchart illustrating yet another method of edge detection, according to an example embodiment;

FIG. 6 is a flow diagram illustrating a type of edge detection in accordance with an exemplary embodiment;

FIG. 7 is a block diagram of a road edge detection apparatus, according to an example embodiment;

fig. 8 is a block diagram of an electronic device, according to an example embodiment.

Detailed Description

Further advantages and effects of the present invention will become readily apparent to those skilled in the art from the disclosure herein, by referring to the accompanying drawings and the preferred embodiments. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be understood that the preferred embodiments are presented by way of illustration only and not by way of limitation.

It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention by way of illustration, and only the components related to the present invention are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, and the form, number and proportion of the components in actual implementation may be arbitrarily changed, and the layout of the components may be more complicated.

In recent years, as lidar has become a popular sensor in the field of autopilot sensing, a related algorithm based on lidar has become a hot spot for industrial research. Among the lidar-based sensing algorithms, the detection of road edge information is an important support for many autopilot backend algorithms. The current main road edge detection algorithm can detect and filter the road edge candidate points based on a single view (front view or top view), or determine the road edge candidate points by utilizing multi-frame laser data under the top view, or determine the road edge candidate points based on the position information and map information of the vehicle, or determine the road edge candidate points by carrying out ground point segmentation on the point cloud data, or detect noise points and invalid points in the point cloud data by a deep learning method and reject the noise points and the invalid points so as to determine the road edge candidate points, and then obtain curve-form output by a curve fitting method.

The method has the defects of detection precision and reliability, and particularly when the road edge candidate points are determined based on the front view method, missed detection is easy to occur when the point cloud distance is far and the point cloud is partially shielded, and the front view gradient algorithm scheme is sensitive to point cloud noise and measurement errors, so that excessive false detection key points are easy to occur. The method for extracting the road edge candidate points based on the top view grids is limited by the constraint of the grid size, and short road edge key points are easy to miss. When the road edge candidate point is determined by utilizing multi-frame laser data under the top view, additional frame pose transformation information is needed to be fused, and part of algorithms need to splice multi-frame point clouds, so that the accuracy of the operation is easily interfered by pose transformation accuracy, and the result performance is obviously affected.

When determining a route edge candidate point based on position information of a vehicle and map information, the map information needs to be acquired in advance, but the map information cannot be acquired in many scenes, resulting in low applicability of the method. When the point cloud data is subjected to ground point segmentation to determine the route edge candidate points, the influence of ground segmentation accuracy is large, and missed detection easily occurs in the process of identifying low-level route edges. When noise points and invalid points in the point cloud data are detected and removed through a deep learning method, the noise points and the invalid points are removed by adopting a deep learning strategy, and the requirement on calculation force is high.

For easy understanding, the following describes the method for detecting the road edge provided in the present application with reference to the accompanying drawings.

The road edge detection method provided by the embodiment of the application can be applied to a road edge detection system. Fig. 1 is a schematic diagram illustrating a construction of a road edge detection system according to an exemplary embodiment. As shown in fig. 1, the road edge detection system 10 includes: the system comprises a data acquisition module 11, a front view data processing module 12, a top view data processing module 13, a road edge key point screening module 14 and a road edge post-processing module 15.

The data acquisition module 11 is configured to acquire a plurality of point cloud data corresponding to a target area; the front view data processing module 12 is configured to construct a front view corresponding to the target area based on the plurality of point cloud data, and determine a gradient value corresponding to each of the plurality of point cloud data based on a coordinate value corresponding to each of the point cloud data in a preset coordinate system of the front view; the top view data processing module 13 is configured to construct a rasterized top view corresponding to the target area based on the plurality of point cloud data, and determine a grid characteristic value of each grid included in the rasterized top view.

The road edge key point screening module 14 is configured to screen at least one point cloud data from the plurality of point cloud data based on a gradient value corresponding to each point cloud data in the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the rasterized top view; the road edge post-processing module 15 performs curve fitting processing based on the screened at least one point cloud data, and determines a road edge position corresponding to the target area, so as to detect the road edge based on the data acquisition module 11, the front view data processing module 12, the top view data processing module 13, the road edge key point screening module 14 and the road edge post-processing module 15.

Fig. 2 is a flowchart illustrating a method of detecting a road edge according to an exemplary embodiment, and as shown in fig. 2, the method of detecting a road edge includes the steps of:

s201, acquiring a plurality of point cloud data corresponding to the target area, and constructing and obtaining a front view corresponding to the target area based on the plurality of point cloud data.

The plurality of point cloud data are data acquired from the target area, and one point cloud data corresponds to one point in the front view.

Optionally, the plurality of point cloud data may be processed to obtain a front view corresponding to the target area. Specifically, the front view corresponding to the target area can be obtained by rasterizing two angles (θ, Φ) in polar coordinates (ρ, θ, Φ) of point cloud data in the original point cloud data, or the front view corresponding to the target area can be obtained by rearranging a plurality of point cloud data in a standard row and column form through a scanning mode of a laser radar.

The plurality of point cloud data corresponding to the target area includes at least one ground point cloud data and at least one non-ground point cloud data.

S202, determining a gradient value corresponding to each point cloud data in a plurality of point cloud data based on coordinate values corresponding to each point cloud data of the front view in a preset coordinate system.

Alternatively, a gradient value corresponding to each point cloud data (excluding the point cloud data in the same column) included in any one column in the front view may be calculated in a preset coordinate system. Specifically, the method for calculating the gradient value corresponding to any one point cloud data P0 (x 0, y0, z 0) is shown in formula one:

wherein d1 is the projection distance of the next point cloud data P1 (X1, Y1, z 1) adjacent to P0 on the same column on the X-Y plane, and d0 is the projection distance of P0 on the X-Y plane.

The calculation method of d1 is shown in the formula II:

the calculation method of d0 is shown in the formula III:

s203, constructing a grid top view corresponding to the target area based on the plurality of point cloud data, and determining grid characteristic values of each grid included in the grid top view.

Wherein the grid characteristic value includes at least one of: the number of point cloud data included in the grid, and the polar coordinate difference value corresponding to the point cloud data included in the grid.

Optionally, the top view may be rasterized on the plurality of point cloud data to obtain a rasterized top view corresponding to the target area, and the point cloud data of each grid included in the rasterized top view is counted to obtain the number of point cloud data included in each grid, the maximum value and the minimum value of x coordinates, the maximum value and the minimum value of y coordinates, and the maximum value and the minimum value of z coordinates corresponding to the point cloud data included in each grid.

The difference value of polar coordinates (i.e., (Zmax-Zmin)) corresponding to the point cloud data included in each grid can be obtained based on the maximum value (i.e., xmax) and the minimum value (i.e., xmin) of the x-coordinate, the maximum value (i.e., ymax) and the minimum value (i.e., ymin) of the y-coordinate, the maximum value (i.e., zmax) and the minimum value (i.e., zmin) of the z-coordinate, and the point cloud data included in each grid.

S204, screening at least one point cloud data from the plurality of point cloud data based on the gradient value corresponding to each point cloud data in the plurality of point cloud data and the grid characteristic value of the grid corresponding to each point cloud data in the grid top view.

S205, performing curve fitting processing based on the screened at least one point cloud data, and determining the road edge position corresponding to the target area.

Optionally, curve fitting may be performed on the at least one selected point cloud data, specifically, a cubic curve may be used to fit the at least one point cloud data, and a final path edge structure in a parameter form is output, or a path edge structure expressed in a key point form is directly output, and further, a path edge position corresponding to the target area may be determined based on the path edge structure.

Fig. 3 is a flowchart illustrating yet another method for detecting a road edge according to an exemplary embodiment, as shown in fig. 3, the method in step S204 described above specifically includes the following steps:

s301, determining a plurality of first point cloud data, corresponding to the point cloud data, of which the gradient value is larger than a preset gradient threshold value, from the plurality of point cloud data.

Alternatively, the plurality of point cloud data having a gradient value greater than the preset gradient threshold value (i.e., the plurality of points P) in each column included in the front view may be determined as the plurality of first point cloud data based on any column included in the front view, searching from the near to the far for the point cloud data greater than the preset gradient threshold value (i.e., the Gth).

S302, determining at least one point cloud data with a grid characteristic value larger than a preset characteristic value of a grid corresponding to the point cloud data in the grid top view from the plurality of first point cloud data.

Optionally, the Grid characteristic value of the Grid may be (Zmax-Zmin), any point P may be projected onto the Grid corresponding to the Grid top view (i.e., grid (P)), and whether the Grid characteristic value of Grid (P) is greater than a preset characteristic value (i.e., zth) is determined, if the Grid characteristic value of Grid (P) is greater than Zth, the point P is an effective key point; otherwise, continuing searching and judging whether the (Zmax-Zmin) of the Grid (P) corresponding to the next point P is larger than Zth or not so as to determine at least one point cloud data which is an effective key point from the plurality of points P.

Fig. 4 is a flowchart illustrating yet another method for detecting a road edge according to an exemplary embodiment, where a difference value of polar coordinates corresponding to point cloud data included in a grid is: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate is shown in fig. 4, and the method further includes the following steps:

s401, determining a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the grid top view based on the coordinate value of each point cloud data in the grid top view.

Alternatively, the maximum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view may be Zmax, and the minimum ordinate value may be Zmin.

S402, determining a polar coordinate difference value corresponding to the point cloud data in each grid on the ordinate based on a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data in each grid in the grid top view.

Alternatively, the point cloud data included in each grid may have a corresponding polar coordinate difference value (Zmax-Zmin) on the ordinate.

Fig. 5 is a flowchart illustrating yet another method for detecting a road edge according to an exemplary embodiment, as shown in fig. 5, the method in step S205 described above specifically includes the following steps:

S501, based on grid characteristic values of grids corresponding to each point cloud data in the screened at least one point cloud data in the grid plan view, clustering the grids in the grid plan view to obtain multiple clusters.

Optionally, based on the screened at least one point cloud data and the grid characteristic value of the grid corresponding to each point cloud data in the non-ground point cloud data obtained through ground detection, performing clustering processing on the grids included in the grid top view through a breadth-first search algorithm to obtain multiple types of clustering clusters.

Specifically, the grid characteristic values may be (Xmax-Xmin), (Ymax-Ymin), and (Zmax-Zmin), when Xgap (i.e., (Xmax-Xmin)), ygap (i.e., (Ymax-Ymin)), and Zgap (i.e., (Zmax-Zmin)) corresponding to any two grids on the x coordinate axis, the y coordinate axis, and the z coordinate axis respectively are all smaller than a preset interval threshold, it is determined that the two grids belong to the same cluster, the same cluster number (i.e., unique code (identity document, ID)) is added to the two grids, and after the cluster numbers are added to all the grids included in the top view of the grid, the clustering process is ended.

It should be noted that, the judgment standard of the same cluster can be adjusted according to different sensor wire harnesses. One type of cluster corresponds to a preset interval threshold, and preset interval thresholds corresponding to different clusters are different. The non-ground point cloud data obtained by ground detection are used for supplementing non-ground point cloud data which are missed in the plurality of point cloud data.

S502, determining a clustering characteristic value corresponding to each type of cluster in the multi-type clusters, and screening at least one type of cluster meeting a preset condition from the multi-type clusters.

Wherein the cluster feature value comprises at least one of: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster, and the preset condition comprises at least one of the following: the clustering characteristic value is larger than the preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value.

Optionally, the corresponding grids in the grid top view may be mapped to polar coordinate grids, cluster feature values corresponding to each type of cluster in the multi-type clusters may be extracted, and at least one type of cluster satisfying the preset condition may be screened from the multi-type clusters based on the cluster feature values corresponding to each type of cluster in the multi-type clusters.

The grid closest to the origin in each column may be selected from a plurality of grids included in each cluster of at least one cluster type among the polar coordinate grids, as an edge grid, to obtain a plurality of edge grids. And then determining a plurality of edge grids with the clustering characteristic value meeting the target preset condition from the edge grids based on the clustering characteristic value corresponding to the clustering cluster where each edge grid in the edge grids is located, and deleting the edge grids which do not meet the target preset condition from the edge grids.

Further, classification is performed based on y values of centroid points corresponding to each of the plurality of edge grids, so that the plurality of edge grids are split into left and right sides, and the centroid points corresponding to each of the plurality of edge grids are output as road edge key points.

It should be noted that at least one type of cluster is a cluster other than a small cluster formed by noise among a plurality of types of clusters. The target preset condition is that the length corresponding to the cluster is within a preset length threshold corresponding to the vehicle and the pedestrian, the width corresponding to the cluster is within a preset width threshold corresponding to the vehicle and the pedestrian, and the height corresponding to the cluster is within a preset height threshold corresponding to the vehicle and the pedestrian.

S503, performing curve fitting processing based on coordinate values corresponding to point cloud data included in at least one type of cluster, and determining a road edge position corresponding to the target area.

Optionally, curve fitting processing may be performed based on coordinate values corresponding to point cloud data (i.e., road edge key points) included in at least one cluster type, so as to determine a road edge position corresponding to the target area.

Fig. 6 is a schematic flow chart of a road edge detection, as shown in fig. 6, in which point cloud data may be input, front views and top views are respectively generated based on the point cloud data, and then gradient values of the point cloud data and grid characteristic values of top view grids of the point cloud data are calculated respectively to determine dual-view key points from the point cloud data. Further, inputting double-view key points and ground-detected non-ground points, updating grid characteristic values of the top view based on the double-view key points and the ground-detected non-ground points, then performing top view grid clustering based on the two-dimensional field to obtain and output top view grid clusters, extracting cluster characteristics of the grid clusters, and filtering discrete grid clusters based on the cluster characteristics.

Inputting a top view grid to generate a polar coordinate grid, reserving an edge grid nearest to the origin of the polar coordinate grid from the filtered grid cluster, removing isolated grids such as vehicles and pedestrians from the edge grid based on cluster characteristics corresponding to the edge grid, and splitting the rest of the edge grid into left and right road edge grids. And finally, curve fitting is carried out on the left and right road edge grids so as to output a road edge curve.

The embodiment of the application provides a road edge detection method, which can be used for extracting preliminary key points through a mode of combining a front view and a top view, detecting the key points of the road edge by combining the advantages of multiple views, solving the detection problem of low road edges and various fences, realizing high extraction precision of the low road edges, reserving the key points formed by the low road edges to the greatest extent, improving the detection precision of the low key points, and solving the problem of insufficient detection precision under a single view. And then merging the extracted preliminary key points with non-ground points to cluster, merging the preliminarily detected key points containing low road edges with the non-ground points to cluster, so that the defect that missed detection possibly exists in the preliminary key points can be overcome, and the performance of road edge detection is improved.

Further, all clustering edge grids containing the road edge grids are extracted through the polar coordinate grids, false detection caused by vehicles and pedestrians is filtered through clustering cluster information under the top view, the top view grids are converted into the polar coordinate grids, then the vehicles and the pedestrians are filtered through clustering cluster features under the top view, further more accurate road edge candidate key points are obtained, the effective distance and the accuracy of road edge detection can be improved through introducing clustering cluster information, and the density of the road edge key points and the detection rate of the road edge key points in a distance can be improved through extracting the edges of the clustering clusters. The non-road-edge targets in the road can be rapidly and effectively eliminated by combining the polar coordinate grid with cluster information, and the problem of false road edge detection caused by vehicles, pedestrians and the like in the road is solved. The method only adopts the traditional machine learning algorithm, has low calculation force requirement on the operation platform, is easy to engineer and land, does not depend on map information and pose transformation information, has strong independence of algorithm modules, and has stronger robustness.

The foregoing description of the solution provided in the embodiments of the present application has been mainly presented in terms of a method. In order to achieve the above functions, the path edge detection device or the electronic device includes a hardware structure and/or a software module that perform respective functions. Those of skill in the art will readily appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be implemented as hardware or combinations of hardware and computer software. Whether a function is implemented as hardware or computer software driven hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

According to the method, the functional modules of the road edge detection device or the electronic device may be divided, for example, the road edge detection device or the electronic device may include each functional module corresponding to each functional division, or two or more functions may be integrated into one processing module. The integrated modules may be implemented in hardware or in software functional modules. It should be noted that, in the embodiment of the present application, the division of the modules is schematic, which is merely a logic function division, and other division manners may be implemented in actual implementation.

Fig. 7 is a block diagram illustrating a road edge detection apparatus according to an example embodiment. Referring to fig. 7, the edge detection device 70 includes: an acquisition module 701, a determination module 702, and a processing module 703;

the acquiring module 701 is configured to acquire a plurality of point cloud data corresponding to a target area, where the plurality of point cloud data is data acquired from the target area;

a processing module 703, configured to construct a front view corresponding to the target area based on a plurality of point cloud data, where one point cloud data corresponds to one point in the front view;

A determining module 702, configured to determine a gradient value corresponding to each of the plurality of point cloud data based on a coordinate value corresponding to each of the point cloud data in the preset coordinate system for the front view;

the processing module 703 is further configured to construct a rasterized top view corresponding to the target area based on the plurality of point cloud data;

the determining module 702 is further configured to determine a grid characteristic value of each grid included in the rasterized top view, where the grid characteristic value includes at least one of: the number of the point cloud data included in the grid and the polar coordinate difference value corresponding to the point cloud data included in the grid;

the processing module 703 is further configured to screen at least one point cloud data from the plurality of point cloud data based on a gradient value corresponding to each point cloud data in the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the grid plan view, and perform curve fitting processing based on the screened at least one point cloud data;

the determining module 702 is further configured to determine a route edge position corresponding to the target area.

In a possible implementation manner, the determining module 702 is further configured to determine, from the plurality of point cloud data, a plurality of first point cloud data with gradient values corresponding to the point cloud data greater than a preset gradient threshold; the determining module 702 is further configured to determine, from the plurality of first point cloud data, at least one point cloud data having a grid characteristic value of a grid corresponding to the point cloud data in the rasterized top view greater than a preset characteristic value.

In one possible implementation, the polar coordinate difference value corresponding to the point cloud data included in the grid is: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate; the determining module 702 is further configured to determine, based on coordinate values of each of the plurality of point cloud data in the rasterized top view, a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each of the grids included in the rasterized top view; the determining module 702 is further configured to determine a polar coordinate difference value corresponding to the point cloud data included in each grid on the ordinate, based on a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view.

In a possible implementation manner, the processing module 703 is further configured to perform clustering processing on the grids included in the top-grid view based on the grid characteristic values of the grids corresponding to each point cloud data in the screened at least one point cloud data in the top-grid view to obtain multiple types of cluster clusters; the determining module 702 is further configured to determine a cluster feature value corresponding to each of the multiple types of clusters, where the cluster feature value includes at least one of: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster; the processing module 703 is further configured to screen at least one cluster type that meets a preset condition from among cluster types, where the preset condition includes at least one of the following: the clustering characteristic value is larger than a preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value; the processing module 703 is further configured to perform curve fitting processing based on coordinate values corresponding to the point cloud data included in at least one type of cluster; the determining module 702 is further configured to determine a route edge position corresponding to the target area.

The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.

Fig. 8 is a block diagram of an electronic device, according to an example embodiment. As shown in fig. 8, electronic devices 80 include, but are not limited to: a processor 801 and a memory 802.

The memory 802 is used for storing executable instructions of the processor 801. It will be appreciated that the processor 801 is configured to execute instructions to implement the method of edge detection in the above embodiments.

It should be noted that the electronic device structure shown in fig. 8 is not limited to the electronic device, and the electronic device may include more or less components than those shown in fig. 8, or may combine some components, or may have different arrangements of components, as will be appreciated by those skilled in the art.

The processor 801 is a control center of the electronic device, connects various parts of the entire electronic device using various interfaces and lines, and performs various functions of the electronic device and processes data by running or executing software programs and/or modules stored in the memory 802, and calling data stored in the memory 802, thereby performing overall monitoring of the electronic device. The processor 801 may include one or more processing modules. Alternatively, the processor 801 may integrate an application processor that primarily handles operating systems, user interfaces, applications, etc., with a modem processor that primarily handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 801.

Memory 802 may be used to store software programs as well as various data. The memory 802 may mainly include a storage program area that may store an operating system, application programs (such as a determination unit, a processing unit, etc.) required for at least one functional module, and a storage data area. In addition, memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.

In an exemplary embodiment, a computer readable storage medium is also provided, e.g., a memory 802, comprising instructions executable by the processor 801 of the electronic device 800 to implement the method of edge detection in the above-described embodiments.

In actual implementation, the functions of the acquisition module 701, the determination module 702, and the processing module 703 in fig. 7 may be implemented by the processor 801 in fig. 8 calling a computer program stored in the memory 802. For specific execution, reference may be made to the description of the path edge detection method in the above embodiment, and details are not repeated here.

Alternatively, the computer readable storage medium may be a non-transitory computer readable storage medium, for example, a read-only memory (ROM), a random access memory (random access memory, RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

In an exemplary embodiment, there is also provided a vehicle including a road edge detection device, by which the road edge detection method in the above-described embodiment can be completed.

In an exemplary embodiment, the present application also provides a computer program product comprising one or more instructions executable by the processor 801 of an electronic device to perform the method of edge detection in the above-described embodiments.

It should be noted that, when the instructions in the computer readable storage medium or one or more instructions in the computer program product are executed by the processor of the electronic device, the processes of the embodiments of the method for detecting a road edge are implemented, and the technical effects same as those of the method for detecting a road edge can be achieved, so that repetition is avoided, and no description is repeated here.

From the foregoing description of the embodiments, it will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of functional modules is illustrated, and in practical application, the above-described functional allocation may be performed by different functional modules according to needs, i.e. the internal structure of the apparatus is divided into different functional modules, so as to perform all the classification parts or part of the functions described above.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of modules or units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another apparatus, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and the parts shown as units may be one physical unit or a plurality of physical units, may be located in one place, or may be distributed in a plurality of different places. The purpose of the embodiment scheme can be achieved by selecting part or all of the classification part units according to actual needs.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a readable storage medium. Based on such understanding, the technical solution of the embodiments of the present application may be essentially or partly contributing to the prior art or the whole classification part or part of the technical solution may be embodied in the form of a software product stored in a storage medium, including several instructions to cause a device (may be a single-chip microcomputer, a chip or the like) or a processor (processor) to perform the whole classification part or part of the steps of the methods of the embodiments of the present application. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, etc.

The foregoing is merely a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope of the present disclosure should be covered in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. The road edge detection method is characterized by comprising the following steps of:

Acquiring a plurality of point cloud data corresponding to a target area, and constructing a front view corresponding to the target area based on the plurality of point cloud data, wherein the plurality of point cloud data are acquired from the target area, and one point cloud data corresponds to one point in the front view;

determining a gradient value corresponding to each point cloud data in the plurality of point cloud data based on a coordinate value corresponding to each point cloud data of the front view in a preset coordinate system;

constructing a gridding top view corresponding to the target area based on the plurality of point cloud data, and determining a grid characteristic value of each grid included in the gridding top view, wherein the grid characteristic value comprises at least one of the following: the number of the point cloud data included in the grid and the polar coordinate difference value corresponding to the point cloud data included in the grid;

and screening at least one point cloud data from the plurality of point cloud data based on a gradient value corresponding to each point cloud data in the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the grid plan view, and performing curve fitting processing based on the screened at least one point cloud data to determine a route edge position corresponding to the target area.

2. The method of claim 1, wherein the screening at least one point cloud data from the plurality of point cloud data based on a gradient value corresponding to each point cloud data of the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the rasterized top view comprises:

determining a plurality of first point cloud data with gradient values larger than a preset gradient threshold value corresponding to the point cloud data from the plurality of point cloud data;

and determining at least one point cloud data of which the grid characteristic value of the grid corresponding to the point cloud data in the grid top view is larger than a preset characteristic value from the plurality of first point cloud data.

3. The method of claim 2, wherein the point cloud data included in the grid corresponds to a polar coordinate difference value of: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate, the method further includes:

determining a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view based on the coordinate value of each point cloud data in the rasterized top view;

And determining a polar coordinate difference value corresponding to the point cloud data included in each grid on the ordinate based on a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view.

4. A method according to any one of claims 1 to 3, wherein the performing curve fitting processing based on the screened at least one point cloud data to determine a route edge position corresponding to the target area includes:

clustering grids included in the gridding top view based on grid characteristic values of grids corresponding to each point cloud data in the at least one point cloud data in the gridding top view to obtain multiple types of clusters;

determining a clustering characteristic value corresponding to each type of cluster in the multi-type clusters, and screening at least one type of cluster meeting a preset condition from the multi-type clusters, wherein the clustering characteristic value comprises at least one of the following: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster, and the preset condition comprises at least one of the following: the clustering characteristic value is larger than a preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value;

And performing curve fitting processing based on coordinate values corresponding to the point cloud data included in the at least one cluster, and determining a road edge position corresponding to the target region.

5. The road edge detection device is characterized by comprising an acquisition module, a determination module and a processing module;

the acquisition module is used for acquiring a plurality of point cloud data corresponding to a target area, wherein the plurality of point cloud data are acquired from the target area;

the processing module is used for constructing and obtaining a front view corresponding to the target area based on the plurality of point cloud data, wherein one point cloud data corresponds to one point in the front view;

the determining module is used for determining a gradient value corresponding to each point cloud data in the plurality of point cloud data based on a coordinate value corresponding to each point cloud data of the front view in a preset coordinate system;

the processing module is further configured to construct a rasterized top view corresponding to the target area based on the plurality of point cloud data;

the determining module is further configured to determine a grid characteristic value of each grid included in the rasterized top view, where the grid characteristic value includes at least one of: the number of the point cloud data included in the grid and the polar coordinate difference value corresponding to the point cloud data included in the grid;

The processing module is further configured to screen at least one point cloud data from the plurality of point cloud data based on a gradient value corresponding to each point cloud data in the plurality of point cloud data and a grid characteristic value of a grid corresponding to each point cloud data in the grid plan view, and perform curve fitting processing based on the screened at least one point cloud data;

the determining module is further configured to determine a route edge position corresponding to the target area.

6. The apparatus according to claim 5, wherein the determining module is further configured to determine, from the plurality of point cloud data, a plurality of first point cloud data having a gradient value corresponding to the point cloud data greater than a preset gradient threshold;

the determining module is further configured to determine, from the plurality of first point cloud data, the at least one point cloud data, where a grid characteristic value of a grid corresponding to the point cloud data in the rasterized top view is greater than a preset characteristic value.

7. The apparatus according to claim 6, wherein the point cloud data included in the grid corresponds to a polar coordinate difference value of: the polar coordinate difference value corresponding to the point cloud data included in the grid on the ordinate;

The determining module is further configured to determine a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view based on the coordinate value of each point cloud data in the rasterized top view;

the determining module is further configured to determine a difference value of polar coordinates corresponding to the point cloud data included in each grid on the ordinate, based on a maximum ordinate value and a minimum ordinate value corresponding to the point cloud data included in each grid included in the rasterized top view.

8. The road edge detection device according to any one of claims 5 to 7, wherein the processing module is further configured to perform clustering processing on grids included in the rasterized top view based on a grid characteristic value of a grid corresponding to each point cloud data in the screened at least one point cloud data in the rasterized top view to obtain multiple types of cluster clusters;

the determining module is further configured to determine a cluster feature value corresponding to each of the multiple types of clusters, where the cluster feature value includes at least one of: the number of the point cloud data included in the cluster, the number of grids included in the cluster, the length corresponding to the cluster, the width corresponding to the cluster and the height corresponding to the cluster;

The processing module is further configured to screen at least one type of cluster that meets a preset condition from the multiple types of clusters, where the preset condition includes at least one of the following: the clustering characteristic value is larger than a preset clustering characteristic value, and the clustering characteristic value is the preset clustering characteristic value;

the processing module is further used for performing curve fitting processing based on coordinate values corresponding to the point cloud data included in the at least one type of cluster;

9. An electronic device, comprising: a processor;

a memory for storing the processor-executable instructions;

wherein the processor is configured to execute the instructions to implement the method of any one of claims 1 to 4.

10. A computer readable storage medium, characterized in that, when computer-executable instructions stored in the computer readable storage medium are executed by a processor of an electronic device, the electronic device is capable of performing the method of any one of claims 1 to 4.

11. A vehicle comprising a road edge detection device according to any one of claims 5 to 8, the vehicle being adapted to implement the method according to any one of claims 1 to 4.