A curb detection method, electronic device and storage medium

Abstract

Embodiments of the present application disclose a road edge detection method, an electronic device and a storage medium. The electronic device obtains road point cloud data; performs edge key point extraction processing based on the road point cloud data to obtain edge key point information corresponding to the road point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one straight line segment information according to the at least one point cluster; wherein the at least one straight line segment information is determined according to a target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing according to the at least one straight line segment information to obtain a road edge detection result.

Description

Technical Field

[0001] This invention relates to the field of automotive technology, and more particularly to a curb detection method, electronic device, and storage medium. Background Technology

[0002] In recent years, LiDAR has become an increasingly common sensor in autonomous driving perception; LiDAR-based recognition algorithms have become a hot research topic in the industry; and in LiDAR-based perception algorithms, the detection of roadside information is an important support for many autonomous driving backend algorithms.

[0003] However, existing LiDAR-based road edge detection algorithms still have some problems. For example, the detection effect is not ideal for some road edges with large curvature. Most algorithms need to obtain road map information in advance, or need to stitch together multiple frames of point clouds, or rely on deep learning to remove interference, which greatly increases the complexity of the algorithm and requires high platform computing power, thus affecting the efficiency of road edge detection. It can be seen that existing road edge detection methods generally suffer from poor detection effect and low detection efficiency. Summary of the Invention

[0004] This application provides a curb detection method, electronic device, and storage medium, which can effectively improve the detection effect and efficiency of curb detection.

[0005] The technical solution of this application embodiment is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a curb detection method, the method comprising:

[0007] Acquire road surface point cloud data;

[0008] Edge key point extraction processing is performed on the road surface point cloud data to obtain the edge key point information corresponding to the road surface point cloud data.

[0009] At least one point cluster is determined based on the edge key point information, and at least one line segment information is determined based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster;

[0010] The curb detection result is obtained by performing aggregation processing based on the information of at least one straight line segment.

[0011] Secondly, embodiments of this application provide an electronic device, which includes an acquisition unit and a processing unit.

[0012] The acquisition unit is used to acquire road surface point cloud data;

[0013] The processing unit is configured to perform edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; and to determine at least one point cluster based on the edge key point information, and to determine at least one line segment information based on the at least one point cluster; wherein the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and to perform aggregation processing based on the at least one line segment information to obtain road edge detection results.

[0014] Thirdly, embodiments of this application provide an electronic device, which includes a processor and a memory storing processor-executable instructions; when the instructions are executed by the processor, the curb detection method described above is implemented.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium having a program stored thereon, which is applied in an electronic device. When the program is executed by a processor, it implements the curb detection method as described above.

[0016] This application provides a curb detection method, an electronic device, and a storage medium. The electronic device acquires road surface point cloud data; performs edge key point extraction processing on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing based on the at least one line segment information to obtain the curb detection result. Therefore, in this application, the electronic device can efficiently extract edge key points from the road surface point cloud data after acquiring it in real time, without needing to obtain map information about the road in advance or stitch the point cloud data before performing subsequent road edge detection. Furthermore, after the electronic device obtains the edge key point information corresponding to the road edge, it can determine at least one point cluster based on the edge key point information, and obtain at least one line segment information based on the target unit vector corresponding to each of the at least one point cluster. Thus, it can perform aggregation processing based on at least one line segment information to obtain the road edge detection result. The overall method can be executed without relying on deep learning algorithms to remove interference, has low complexity, low computational requirements, and can effectively improve the detection effect and efficiency of road edge detection. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 1 ;

[0018] Figure 2 This is a schematic diagram illustrating the edge key point extraction proposed in an embodiment of this application;

[0019] Figure 3 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 2 ;

[0020] Figure 4 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 3 ;

[0021] Figure 5 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 4 ;

[0022] Figure 6 This is a schematic diagram of the composition structure of the electronic device proposed in the embodiments of this application. Figure 1 ;

[0023] Figure 7 This is a schematic diagram of the composition structure of the electronic device proposed in the embodiments of this application. Figure 2 . Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the relevant application and not for limiting the application. Furthermore, it should be noted that, for ease of description, only the parts related to the relevant application are shown in the accompanying drawings.

[0025] In recent years, LiDAR has become an increasingly common sensor in autonomous driving perception; LiDAR-based recognition algorithms have become a hot research topic in the industry; and in LiDAR-based perception algorithms, the detection of roadside information is an important support for many autonomous driving backend algorithms. Most current road edge detection algorithms are based on a single view, detecting and filtering candidate points, and then obtaining a curve output through curve fitting. However, these methods suffer from shortcomings in both detection accuracy and reliability. Furthermore, many laser point cloud-based road edge detection algorithms often fail to accurately fit the actual road edge when it has high curvature, such as when the road edge contains right-angle turns. Some algorithms also have many unmet prerequisites, such as requiring prior access to road map information, which is often difficult to achieve. Other algorithms require stitching together multiple point cloud frames, the accuracy of which is easily affected by pose transformation accuracy. Some algorithms rely on deep learning to remove interference, increasing complexity. Still others require multiple curve fittings based on RANSAC iterative algorithms, but in practical deployments, chip computing power often limits real-time performance, impacting road edge detection efficiency. Therefore, existing road edge detection methods generally suffer from poor detection performance and low efficiency.

[0026] To address the problems existing in the prior art, this application provides a curb detection method, electronic device, and storage medium. The electronic device acquires road surface point cloud data; performs edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing based on the at least one line segment information to obtain the curb detection result. This method can effectively improve the detection effect and efficiency of curb detection.

[0027] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0028] This application provides a curb detection method. Figure 1 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 1 ,like Figure 1 As shown, the curb detection method for electronic devices may include the following steps:

[0029] Step 101: Obtain road surface point cloud data.

[0030] In the embodiments of this application, the electronic device may first acquire road surface point cloud data.

[0031] It should be noted that, in the embodiments of this application, LiDAR can be used for real-time scanning to obtain laser data of the road surface, and then road surface point cloud data can be obtained based on the laser data; that is, the road surface point cloud data is point cloud data obtained based on the real-time condition of the road surface, which contains real-time road surface information.

[0032] It should be noted that, in the embodiments of this application, the electronic device can be any device with communication and storage functions, and the electronic device can have resource selection requirements. For example, the electronic device can be an in-vehicle electronic device, a smart lock, a tablet computer, a mobile phone, an e-reader, a remote control, a personal computer (PC), a laptop computer, a smart TV, a wearable electronic device, a personal digital assistant (PDA), a portable media player (PMP), a navigation device, and other electronic devices.

[0033] Step 102: Extract edge key points based on road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data.

[0034] In the embodiments of this application, after acquiring road surface point cloud data, the electronic device can perform edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data.

[0035] It should be noted that, in the embodiments of this application, the edge key point information is the information corresponding to the curb, so that subsequent processing can be performed based on the edge key point information to obtain the detection result of the curb.

[0036] Furthermore, in the embodiments of this application, when obtaining edge key point information, it is necessary to filter the ground points in the input road surface point cloud data to obtain the edge key points.

[0037] Furthermore, in some embodiments of this application, when the electronic device performs edge key point extraction processing based on road point cloud data to obtain edge key point information corresponding to the road point cloud data, it can first preprocess the road point cloud data to obtain front view information corresponding to the road point cloud data; and then perform edge key point extraction processing based on the front view information to obtain edge key point information.

[0038] It should be noted that, in the embodiments of this application, the preprocessing of the road surface point cloud data is mainly to remove ground points and obtain a front view, so that edge key points can be extracted quickly based on the front view.

[0039] Furthermore, in some embodiments of this application, the preprocessing can be polar coordinate rasterization or rearrangement; wherein, polar coordinate rasterization is rasterization based on the polar coordinate information of points in the road surface point cloud data; and rearrangement is rearranging the road surface point cloud data into a row and column format.

[0040] Furthermore, in the embodiments of this application, for the method of obtaining the front view through polar coordinate rasterization processing, for example, the polar coordinate information of points in the road surface point cloud data is represented as follows: Where ρ represents the distance of the point from the origin, and θ represents the horizontal angle. Represents the vertical coordinate value; θ can be used. These two values ​​enable the rasterization of road surface point cloud data to obtain a front view.

[0041] Furthermore, in the embodiments of this application, for the method of rearranging the processing to obtain the front view, for example, the scanning mode of the LiDAR can be used to rearrange the road point cloud data into a standardized row and column form to obtain the front view.

[0042] Furthermore, in the embodiments of this application, when performing edge key point extraction processing based on the front view information to obtain edge key point information, for example, the edge grid under the top view can be extracted based on the front view information; for example, in the front view obtained by polar coordinate rasterization processing, the grid closest to the origin in each column can be taken as the edge grid. Figure 2 This is a schematic diagram illustrating the edge key point extraction proposed in an embodiment of this application, as shown below. Figure 2 As shown, white represents the grid cells occupied by ground points, while edge grid cells and other grid cells are the remaining grid cells after removing ground points. By taking the bottom grid cell in each column, all the outermost grid cells can be quickly found. For example, when the algorithm complexity for searching edge grid cells is O(N), where N is the total number of grid cells, the points falling in the grid cells are the edge key points, thus obtaining the edge key point information.

[0043] It should be noted that, in the embodiments of this application, the front view is obtained by polar coordinate rasterization or rearrangement, and candidate edge key points are extracted based on the front view, which can greatly reduce the search space for subsequent line fitting and improve the running speed.

[0044] Furthermore, in embodiments of this application, the electronic device may include an edge key point extraction module for acquiring the aforementioned edge key point information.

[0045] Step 103: Determine at least one point cluster based on edge key point information, and determine at least one line segment information based on at least one point cluster; wherein, at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster.

[0046] In the embodiments of this application, after the electronic device performs edge key point extraction processing based on road point cloud data to obtain edge key point information corresponding to the road point cloud data, it can determine at least one point cluster based on the edge key point information and determine at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster.

[0047] It should be noted that, in the embodiments of this application, by means of line segment fitting processing, at least one line segment information that can fit these edge key points can be quickly detected.

[0048] Furthermore, in some embodiments of this application, when the electronic device determines at least one point cluster based on edge key point information and determines at least one line segment information based on at least one point cluster, it may first sort the edge key point information to obtain sorted key point information; then determine at least one point cluster based on the sorted key point information; and then perform line segment fitting processing based on at least one point cluster to obtain at least one line segment information corresponding to at least one point cluster.

[0049] Furthermore, in some embodiments of this application, when the electronic device sorts the edge key point information to obtain sorted key point information, it can first determine the horizontal angle information corresponding to each edge key point in the edge key point information; then sort the edge key points in the edge key point information according to the horizontal angle information to obtain sorted key point information.

[0050] It should be noted that, in the embodiments of this application, when sorting the edge key points in the edge key point information according to the horizontal angle information, the horizontal angle information can be arranged in ascending order, and then the sorted key point information can be obtained according to the edge key points corresponding to the arranged horizontal angle information; or the horizontal angle information can be arranged in descending order, and then the sorted key point information can be obtained according to the edge key points corresponding to the arranged horizontal angle information.

[0051] For example, in an embodiment of this application, when sorting three edge key points A1, A2, and A3, the horizontal angle information corresponding to each of the three edge key points is θ1, θ2, and θ3, respectively. The result of arranging the horizontal angle information in ascending order is θ1 < θ3 < θ2, so the sorted key point information is A1, A3, and A2; while when arranging the horizontal angle information in descending order, the result is θ2 > θ3 > θ1, so the sorted key point information is A2, A3, and A1.

[0052] Furthermore, in the embodiments of this application, after obtaining the sorted key point information, the sorted key point information can be divided into at least one point cluster, and these point clusters can constitute a point cluster set.

[0053] Furthermore, in the embodiments of this application, a point cluster can be selected from this point cluster set for line segment fitting to obtain the line segment information corresponding to this point cluster. Then, another point cluster can be selected and the same step can be performed. This process is repeated until the point cluster set is empty, thus completing the acquisition of the line segment information corresponding to all point clusters in the point cluster set and obtaining at least one line segment information corresponding to at least one point cluster.

[0054] Furthermore, in some embodiments of this application, when the electronic device determines at least one point cluster based on the sorted key point information, it may first obtain the first distance information between every two adjacent edge key points in the sorted key point information; thereby determining at least one point cluster based on the first distance information and the first distance threshold.

[0055] In some embodiments of this application, edge key points corresponding to the first distance information that are less than or equal to the first distance threshold can be divided into a point cluster by comparing the first distance information and the first distance threshold.

[0056] For example, in the embodiments of this application, the sorted key points are A1, A2, A3, A4, A5, A6, and A7. The first distance information between A1 and A2 is 1, the distance information between A2 and A3 is 1, the distance information between A3 and A4 is 4, the distance information between A4 and A5 is 1, the distance information between A5 and A6 is 3, and the distance information between A6 and A7 is 1. The first distance threshold is 2, so the sorted key points can be divided into 3 point clusters, namely point clusters B1, B2, and B3. Among them, the edge key points in B1 include A1, A2, and A3, the edge key points in B2 include A4 and A5, and the edge key points in B3 include A6 and A7.

[0057] It should be noted that, in the embodiments of this application, the size of the first distance information is not specifically limited.

[0058] Furthermore, in some embodiments of this application, when the electronic device performs line segment fitting processing based on at least one point cluster to obtain at least one line segment information corresponding to at least one point cluster, for one point cluster in the at least one point cluster, it can first determine at least one unit vector corresponding to the point cluster; then determine the target unit vector corresponding to the point cluster based on the at least one unit vector; and then determine the line segment information corresponding to the point cluster based on the target unit vector, so as to obtain at least one line segment information corresponding to at least one point cluster.

[0059] It is understood that in the embodiments of this application, the target unit vector is a unit vector selected from at least one unit vector corresponding to a point cluster. Thus, for each point cluster in the set of point clusters, a target unit vector can be selected. Then, based on the target unit vector corresponding to each point cluster, the information of a line segment corresponding to that point cluster can be determined, thereby obtaining at least one line segment information corresponding to at least one point cluster in the set of point clusters.

[0060] It should be noted that, in the embodiments of this application, the line segment information is determined based on the target unit vector and the inline point corresponding to the target unit vector; wherein, the target unit vector can determine the direction of the line segment, and the inline point can determine the specific length of the line segment, so that a line segment information can be determined together based on the target unit vector and the inline point corresponding to the target unit vector.

[0061] Furthermore, in some embodiments of this application, when determining at least one unit vector corresponding to a point cluster, the electronic device can determine a unit vector based on every two adjacent edge key points in a point cluster to obtain at least one unit vector.

[0062] For example, in an embodiment of this application, assuming there are L points in a point cluster, a unit vector can be determined based on every two adjacent edge key points in the point cluster, thereby obtaining L-1 unit vectors.

[0063] Furthermore, in some embodiments of this application, when the electronic device determines the target unit vector corresponding to a point cluster based on at least one unit vector, it can calculate the second distance information between each unit vector in the at least one unit vector and each edge key point in the point cluster; then determine the inline point information corresponding to each unit vector in the at least one unit vector based on the second distance information; and then determine the target unit vector based on the inline point information corresponding to each unit vector.

[0064] For example, in an embodiment of this application, for L-1 unit vectors corresponding to a point cluster, the second distance information between each unit vector and all edge key points in the point cluster can be calculated.

[0065] It should be noted that, in the embodiments of this application, the inline point is the same as the inline point.

[0066] Furthermore, in the embodiments of this application, the second distance information can be Euclidean distance.

[0067] Furthermore, in the embodiments of this application, the inline point information includes the number of inline points and the line segment length information; wherein, the line segment length information is the length information of the line segment formed by the inline points.

[0068] Furthermore, in some embodiments of this application, when the electronic device determines the target unit vector based on the inline point information corresponding to each unit vector, it can determine the target unit vector according to the number of inline points and the length of the line segment.

[0069] Furthermore, in some embodiments of this application, when the electronic device determines the inline point information corresponding to each unit vector in at least one unit vector based on the second distance information, for one unit vector in at least one unit vector, the edge key point corresponding to the distance information less than the second distance threshold in the second distance information corresponding to the unit vector can be taken as an inline point corresponding to the unit vector, so as to obtain the inline point information corresponding to each unit vector in at least one unit vector.

[0070] For example, in an embodiment of this application, the second distance threshold is denoted as e. The edge key points corresponding to the distance information less than e in the second distance information corresponding to a unit vector can be used as the inline points corresponding to this unit vector. Furthermore, the number of inline points of this unit vector and the length information of the straight line segments formed by the inline points of this unit vector can be counted.

[0071] Furthermore, in some embodiments of this application, when determining the target unit vector based on the number of inline points and the length of the line segment, the first calculation result corresponding to each of the at least one unit vector can be obtained based on the number of inline points and the length of the line segment corresponding to each of the at least one unit vector, as well as the first weight information, and the unit vector corresponding to the largest calculation result among the first calculation results is taken as the target unit vector.

[0072] It should be noted that, in the embodiments of this application, the first weight information can be freely set according to the actual situation, and its size is not specifically limited in this application.

[0073] For example, in an embodiment of this application, the first weight information includes α and β. When calculating the first calculation result for a unit vector, the number of inline points corresponding to the unit vector is 3, and the length information of the line segment formed by its inline points is 9. Then the first calculation result can be α×3+β×9.

[0074] Furthermore, in the embodiments of this application, after determining the target unit vector corresponding to the point cluster, the inline points corresponding to the target unit vector can be moved out of the point cluster.

[0075] Furthermore, in the embodiments of this application, after removing the inline point corresponding to the target unit vector from the point cluster, two new point clusters can be formed based on the points before the starting point and the points after the ending point of the straight line segment formed by the inline point corresponding to the target unit vector. According to a preset point count threshold, the edge key points in the point clusters with a point count greater than or equal to the preset point count threshold are put back into the point cluster set, and the edge key points in the point clusters with a point count less than the preset point count threshold are discarded.

[0076] It should be noted that, in the embodiments of this application, the above-mentioned straight line segment fitting process can quickly achieve the fitting of multiple straight line segments. This method can run in real time on embedded platforms with low computing power, has low computing power requirements, and thus can improve the efficiency of curb detection.

[0077] It should be noted that, in the embodiments of this application, after obtaining at least one line segment information, subsequent processing can be performed based on at least one line segment information to obtain the final curb detection result; this application uses the form of multiple line segments to describe the curb, which can conveniently represent different types of curb conditions, without being limited by the curb direction and curvature.

[0078] Furthermore, in embodiments of this application, the electronic device may include a line segment fitting algorithm module for acquiring at least one line segment information corresponding to the aforementioned edge key point information.

[0079] Step 104: Perform aggregation processing based on at least one straight line segment information to obtain the curb detection result.

[0080] In the embodiments of this application, after the electronic device determines at least one point cluster based on edge key point information and determines at least one straight line segment information based on at least one point cluster, it can perform aggregation processing based on at least one straight line segment information to obtain the curb detection result.

[0081] It should be noted that, in the embodiments of this application, similar straight segments in at least one straight segment information can be aggregated using an aggregation algorithm to obtain the curb detection result.

[0082] Furthermore, in some embodiments of this application, when the electronic device performs aggregation processing based on at least one straight line segment information to obtain a curb detection result, it can determine the target straight line segment information among the at least one straight line segment information whose similarity information is greater than or equal to a similarity threshold; then it performs aggregation processing on the target straight line segment information to obtain a curb detection result.

[0083] Furthermore, in the embodiments of this application, the similarity measurement standard may include minimum distance information between line segments and angle information between line segments; wherein, the minimum distance information between line segments is the minimum distance between the endpoints of two line segments, and the angle information between line segments is the cosine value of the angle between two line segments.

[0084] Furthermore, in the embodiments of this application, the minimum distance information of a line segment can be the minimum value among the four distance information formed by the four endpoints of two line segments.

[0085] Furthermore, in the embodiments of this application, similarity information can be calculated based on the second weight information, the minimum distance information of the line segments, and the angle information of the line segments. Then, two line segments with similarity information greater than the similarity threshold are taken as target line segments, and the target line segments are aggregated to obtain the curb detection result corresponding to the target line segments.

[0086] Furthermore, in the embodiments of this application, after determining the target line segment information with similarity, outlier line segments can be filtered out, wherein an outlier line segment is a line segment that does not have any similarity with any other line segment.

[0087] Furthermore, in the embodiments of this application, after determining the target line segment information, the target line segment information can be assigned the same aggregation identifier, so that the electronic device can perform aggregation processing based on the target line segment information with the same aggregation identifier.

[0088] Furthermore, in the embodiments of this application, after aggregating the pairwise line segment information with similarity in at least one line segment information, i.e. all the target line segment information, a set composed of line segments can be obtained. This set is the curb detection result, and each line segment in the curb detection result can represent a possible curb.

[0089] Furthermore, in the embodiments of this application, after obtaining the curb detection results, the curb detection results can be filtered to further improve the accuracy of curb detection.

[0090] Furthermore, in embodiments of this application, the electronic device may include a straight line aggregation and filtering module for performing the aggregation processing of the aforementioned straight line segment information and the filtering processing of the curb detection results.

[0091] This application provides a curb detection method, in which an electronic device acquires road surface point cloud data; performs edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing based on the at least one line segment information to obtain the curb detection result. Therefore, in this application, the electronic device can efficiently extract edge key points from the road surface point cloud data after acquiring it in real time, without needing to obtain map information about the road in advance or stitch the point cloud data before performing subsequent road edge detection. Furthermore, after the electronic device obtains the edge key point information corresponding to the road edge, it can determine at least one point cluster based on the edge key point information, and obtain at least one line segment information based on the target unit vector corresponding to each of the at least one point cluster. Thus, it can perform aggregation processing based on at least one line segment information to obtain the road edge detection result. The overall method can be executed without relying on deep learning algorithms to remove interference, has low complexity, low computational requirements, and can effectively improve the detection effect and efficiency of road edge detection.

[0092] Based on the above embodiments, in another embodiment of this application, the electronic device is, for example, an in-vehicle electronic device; Figure 3 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 2 ,like Figure 3 As shown, the vehicle-mounted electronic device can first acquire ground data to be processed (step 201). The ground data to be processed can be point cloud data or other ground detection results that are not point cloud data. Then, the input ground data to be processed is transformed into a front view, and edge key points are extracted based on the front view to obtain edge key point information (step 202). Then, line segment fitting processing is performed based on the edge key point information to obtain multiple line segment information (step 203). Then, aggregation processing is performed based on the multiple line segment information to obtain the curb detection result (step 204). Furthermore, filtering processing can be performed based on the curb detection result to obtain the filtered curb detection result (step 205).

[0093] Furthermore, in the embodiments of this application, a front view can be obtained by filtering ground points on the point cloud data; the specific method for obtaining the front view can be polar coordinate rasterization or rearrangement processing.

[0094] Furthermore, in the embodiments of this application, for the method of obtaining the front view through polar coordinate rasterization processing, for example, the polar coordinate information of points in the road surface point cloud data is represented as follows: Where ρ represents the distance of the point from the origin, and θ represents the horizontal angle. Represents the vertical coordinate value; θ can be used. These two values ​​enable the rasterization of road surface point cloud data to obtain a front view.

[0095] Furthermore, in the embodiments of this application, for the method of rearranging the processing to obtain the front view, for example, the scanning mode of the LiDAR can be used to rearrange the road point cloud data into a standardized row and column form to obtain the front view.

[0096] Furthermore, in the embodiments of this application, when performing edge key point extraction processing based on the front view information to obtain edge key point information, for example, the edge grid under the top view can be extracted based on the front view information; for example, in the front view obtained by polar coordinate rasterization processing, the grid closest to the origin in each column can be taken as the edge grid, such as... Figure 2 As shown, white represents the grid cells occupied by ground points, while edge grid cells and other grid cells are the remaining grid cells after removing ground points. By taking the bottom grid cell in each column, all the outermost grid cells can be quickly found. For example, when the algorithm complexity for searching edge grid cells is O(N), where N is the total number of grid cells, the points falling in the grid cells are the edge key points, thus obtaining the edge key point information.

[0097] Furthermore, in the embodiments of this application, the vehicle-mounted electronic device can utilize the edge key point extraction module to acquire the aforementioned edge key point information.

[0098] Furthermore, in the embodiments of this application, the vehicle-mounted electronic device can use a straight line segment fitting algorithm module to obtain at least one straight line segment information corresponding to the edge key point information; Figure 4 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 3 ,like Figure 4As shown, when performing line segment fitting, the edge key point information can be sorted first (step 301), then the sorted key point information can be split into at least one point cluster, and a point cluster set can be formed based on at least one point cluster (step 302). Then, it can be determined whether the point cluster set is empty (step 303). When it is determined that the point cluster set is not empty, a point cluster is taken out (step 304). Then, at least one unit vector corresponding to the point cluster is determined (step 305), and then the target unit vector is determined based on at least one unit vector (step 306). The line segment information corresponding to the point cluster is determined by measuring the number of points. The points in the point cluster, except for the inline points corresponding to the target unit vector, are split into two new point clusters, one at the head and one at the tail (step 307). Then, it is determined whether the number of points in the two new point clusters is greater than or equal to the preset point number threshold (step 308). Then, the edge key points in the point clusters with the number of points greater than or equal to the preset point number threshold are put back into the point cluster set, and the edge key points in the point clusters with the number of points less than the preset point number threshold are discarded (step 309). The line segment fitting ends when the point cluster set is empty (step 310).

[0099] Furthermore, in the embodiments of this application, when the vehicle-mounted electronic device obtains at least one line segment information corresponding to the edge key point information using the line segment fitting algorithm module, it can first sort the edge key points in the edge key point information according to the horizontal angle information. Specifically, the horizontal angle information can be arranged in ascending order, and then the sorted key point information can be obtained according to the edge key points corresponding to the arranged horizontal angle information; or the horizontal angle information can be arranged in descending order, and then the sorted key point information can be obtained according to the edge key points corresponding to the arranged horizontal angle information.

[0100] Furthermore, in the embodiments of this application, after obtaining the sorted key point information, the sorted key point information can be divided into at least one point cluster, and these point clusters can constitute a point cluster set Θ.

[0101] Furthermore, in the embodiments of this application, a point cluster can be selected from the point cluster set Θ and subjected to line segment fitting processing to obtain the line segment information corresponding to the point cluster. Then, another point cluster can be selected and the same step can be performed. This process is repeated until the point cluster set is empty, thus completing the acquisition of the line segment information corresponding to all point clusters in the point cluster set and obtaining at least one line segment information corresponding to at least one point cluster.

[0102] Further, in the embodiments of this application, when determining the line segment information corresponding to a point cluster, for example, assuming there are L points in the point cluster, a unit vector can be determined based on every two adjacent edge keypoints in the point cluster, thereby obtaining L-1 unit vectors; for the L-1 unit vectors corresponding to the point cluster, the second distance information between each unit vector and all edge keypoints in the point cluster can be calculated respectively; the second distance threshold is denoted as e, and the edge keypoints corresponding to the distance information less than e in the second distance information corresponding to a unit vector can be taken as the inline points corresponding to this unit vector; furthermore, the number of inline points of this unit vector and the length information of the line segment formed by the inline points of this unit vector can be counted. Furthermore, by combining the number of inline points and the length of the line segment formed by the inline points of this unit vector, the best unit vector in this point cluster, i.e., the target unit vector, can be selected. The line segment corresponding to this point cluster can be determined based on the target unit vector and the inline points corresponding to the target unit vector. Then, the inline points corresponding to the target unit vector can be removed from the point cluster, and two new point clusters can be formed by the points before the starting point and the points after the ending point of the line segment formed by the inline points corresponding to the target unit vector. According to the preset point number threshold, the edge key points in the point clusters with a point number greater than or equal to the preset point number threshold are put back into the point cluster set Θ, and the edge key points in the point clusters with a point number less than the preset point number threshold are discarded.

[0103] Furthermore, in the embodiments of this application, the vehicle-mounted electronic device can use the line aggregation and filtering module to aggregate and filter at least one line segment information obtained by the line segment fitting algorithm module.

[0104] Furthermore, in the embodiments of this application, the aggregation process mainly aggregates line segments by comparing the similarity of two line segments. The similarity measurement standard used may include the minimum distance information of the line segments and the angle information between the line segments. The similarity information between each line segment can be determined by combining these two pieces of information.

[0105] For example, in the embodiments of this application, Figure 5 This is a schematic diagram of the implementation process of the curb detection method proposed in the embodiments of this application. Figure 4 ,like Figure 5As shown, at least one line segment information can be obtained first (step 401), then the next line segment is taken out (step 402), and it is determined whether it has been aggregated (step 403). If it has been aggregated, the next line segment is taken out, and step 402 is continued. If it has not been aggregated, the similarity information between the line segment and another first line segment is obtained (step 404), and it is determined whether the similarity information is greater than or equal to the similarity threshold (step 405). If it is, the two line segments are used as target line segment information and assigned the same aggregation identifier for aggregation processing (step 406). If the similarity information between the line segment and the first line segment is less than the similarity threshold, step 402 is continued, and the next line segment information is taken out, for example, a second line segment different from the first line segment is taken out for comparison of the above similarity information, until the comparison of the similarity information of all line segments in at least one line segment information is completed.

[0106] Furthermore, in the embodiments of this application, after the aggregation process is completed, a set consisting of line segments is formed, which is the curb detection result. Each line segment in the curb detection result can represent a possible curb.

[0107] Furthermore, in the embodiments of this application, after obtaining the curb detection results, the curb detection results can be filtered by the straight line aggregation and filtering module, thereby further improving the accuracy of curb detection.

[0108] In summary, in the embodiments of this application, the roadside is represented as a description of multiple straight line segments, and fast multi-straight line segment fitting can be achieved, thereby enabling fast roadside detection; it can also achieve fast extraction of edge key points; by sorting the edge key point information and using the distance between adjacent points to split the edge key points into different point clusters, the processing speed of subsequent straight line segment fitting can be improved, thereby enabling the best straight line fitting corresponding to a point cluster to be quickly determined.

[0109] This application provides a curb detection method, in which an electronic device acquires road surface point cloud data; performs edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing based on the at least one line segment information to obtain the curb detection result. Therefore, in this application, the electronic device can efficiently extract edge key points from the road surface point cloud data after acquiring it in real time, without needing to obtain map information about the road in advance or stitch the point cloud data before performing subsequent road edge detection. Furthermore, after the electronic device obtains the edge key point information corresponding to the road edge, it can determine at least one point cluster based on the edge key point information, and obtain at least one line segment information based on the target unit vector corresponding to each of the at least one point cluster. Thus, it can perform aggregation processing based on at least one line segment information to obtain the road edge detection result. The overall method can be executed without relying on deep learning algorithms to remove interference, has low complexity, low computational requirements, and can effectively improve the detection effect and efficiency of road edge detection.

[0110] Based on the above embodiments, in another embodiment of this application... Figure 6 This is a schematic diagram of the composition structure of the electronic device proposed in the embodiments of this application. Figure 1 ,like Figure 6 As shown, the electronic device 10 proposed in this application embodiment may include: an acquisition unit 11 and a processing unit 12.

[0111] The acquisition unit 11 is used to acquire road surface point cloud data.

[0112] The processing unit 12 is configured to perform edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; and to determine at least one point cluster based on the edge key point information, and to determine at least one line segment information based on the at least one point cluster; wherein the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and to perform aggregation processing based on the at least one line segment information to obtain road edge detection results.

[0113] Furthermore, the processing unit 12 is also used to preprocess the road surface point cloud data to obtain the front view information corresponding to the road surface point cloud data; and to perform edge key point extraction processing based on the front view information to obtain the edge key point information.

[0114] Furthermore, the preprocessing is polar coordinate rasterization or rearrangement; wherein, the polar coordinate rasterization is rasterization based on the polar coordinate information of the points in the road surface point cloud data; and the rearrangement is rearranging the road surface point cloud data into a row and column format.

[0115] Furthermore, the processing unit 12 is also configured to sort the edge key point information to obtain sorted key point information; determine at least one point cluster based on the sorted key point information; and perform line segment fitting processing based on the at least one point cluster to obtain at least one line segment information corresponding to the at least one point cluster.

[0116] Furthermore, the processing unit 12 is also configured to, for one of the point clusters in the at least one point cluster, determine at least one unit vector corresponding to the point cluster; and determine a target unit vector corresponding to the point cluster based on the at least one unit vector; and determine line segment information corresponding to the point cluster based on the target unit vector, so as to obtain at least one line segment information corresponding to the at least one point cluster.

[0117] Furthermore, the processing unit 12 is also configured to obtain first distance information between every two adjacent edge key points in the sorted key point information; and to determine the at least one point cluster based on the first distance information and the first distance threshold.

[0118] Furthermore, the processing unit 12 is also used to determine the horizontal angle information corresponding to each edge key point in the edge key point information; and to sort the edge key points in the edge key point information according to the horizontal angle information to obtain the sorted key point information.

[0119] Furthermore, the processing unit 12 is also configured to determine a unit vector based on every two adjacent edge key points in the point cluster to obtain the at least one unit vector.

[0120] Furthermore, the processing unit 12 is also configured to calculate second distance information between each unit vector in the at least one unit vector and each edge key point in the point cluster; and to determine inline point information corresponding to each unit vector in the at least one unit vector based on the second distance information; and to determine the target unit vector based on the inline point information corresponding to each unit vector.

[0121] Furthermore, the inline point information includes the number of inline points and the line segment length information; wherein, the line segment length information is the length information of the line segment formed by the inline points; the processing unit 12 is also used to determine the target unit vector based on the number of inline points and the line segment length information.

[0122] Furthermore, the processing unit 12 is also configured to, for one of the at least one unit vectors, take the edge key points corresponding to the distance information less than the second distance threshold in the second distance information corresponding to the unit vector as the inline points corresponding to the unit vector, so as to obtain the inline point information corresponding to each of the at least one unit vectors.

[0123] Further, the processing unit 12 determines the target straight segment information whose similarity information is greater than the similarity threshold among the at least one straight segment information; and performs aggregation processing on the target straight segment information to obtain the curb detection result.

[0124] In the embodiments of this application, further, Figure 7 This is a schematic diagram of the composition structure of the electronic device proposed in the embodiments of this application. Figure 2 ,like Figure 7 As shown, the electronic device 10 proposed in this application embodiment may further include a processor 13 and a memory 14 storing instructions executable by the processor 13; furthermore, the electronic device 10 may also include a communication interface 15 and a bus 16 for connecting the processor 13, the memory 14 and the communication interface 15.

[0125] In the embodiments of this application, the processor 13 can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), electronic device, microelectronic device, and microprocessor. It is understood that for different devices, the electronic device used to implement the above-mentioned processor function can also be other types, and this application embodiment does not specifically limit this. The electronic device 10 may also include a memory 14, which can be connected to the processor 13. The memory 14 is used to store executable program code, which includes computer operation instructions. The memory 14 may include high-speed RAM memory and may also include non-volatile memory, such as at least two disk drives.

[0126] In embodiments of this application, bus 16 is used to connect communication interface 15, processor 13 and memory 14 and the mutual communication between these devices.

[0127] In embodiments of this application, memory 14 is used to store instructions and data.

[0128] Further, in an embodiment of this application, the processor 13 is configured to acquire road surface point cloud data; perform edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determine at least one point cluster based on the edge key point information, and determine at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and perform aggregation processing based on the at least one line segment information to obtain road edge detection results.

[0129] In practical applications, the aforementioned memory 14 can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or a combination of the above types of memory, and provide instructions and data to the processor 13.

[0130] Furthermore, in this embodiment, the functional modules can be integrated into one analysis unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0131] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0132] This application provides an electronic device that acquires road surface point cloud data; performs edge key point extraction processing on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; determines at least one point cluster based on the edge key point information, and determines at least one line segment information based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster; and performs aggregation processing based on the at least one line segment information to obtain a road edge detection result. Therefore, in this application, the electronic device can efficiently extract edge key points from the road surface point cloud data after real-time acquisition, without needing to obtain map information about the road in advance or stitch the point cloud data before performing subsequent road edge detection; furthermore, after obtaining the edge key point information corresponding to the road edge, the electronic device can determine at least one point cluster based on the edge key point information, and obtain at least one line segment information based on the target unit vector corresponding to each of the at least one point cluster, thereby enabling aggregation processing based on the at least one line segment information to obtain a road edge detection result; the overall method can be executed without relying on deep learning algorithms to remove interference, has low complexity, low computational requirements, and can effectively improve the detection effect and efficiency of road edge detection.

[0133] Specifically, the program instructions corresponding to a curb detection method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives; when the program instructions corresponding to a curb detection method in the storage media are read or executed by an electronic device, the following steps are included:

[0134] Acquire road surface point cloud data;

[0135] Edge key point extraction processing is performed on the road surface point cloud data to obtain the edge key point information corresponding to the road surface point cloud data.

[0136] At least one point cluster is determined based on the edge key point information, and at least one line segment information is determined based on the at least one point cluster; wherein, the at least one line segment information is determined based on the target unit vector corresponding to each of the at least one point cluster;

[0137] The curb detection result is obtained by performing aggregation processing based on the information of at least one straight line segment.

[0138] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0139] This application is described with reference to schematic and / or block diagrams of implementations of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the schematic and / or block diagrams can be implemented by computer program instructions, and combinations of blocks in the schematic and / or block diagrams can be implemented. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the schematic and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0140] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in the implementation flow diagram. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0141] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0142] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application.

Claims

1. A curb detection method characterized by comprising: The method comprises: obtaining road surface point cloud data; performing edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data; performing sorting processing on the edge key point information to obtain sorted key point information; determining at least one point cluster according to the sorted key point information; determining a unit vector according to each two adjacent edge key points in the one point cluster to obtain at least one unit vector; calculating second distance information between each unit vector in the at least one unit vector and each edge key point in the one point cluster respectively; determining in-line point information corresponding to each unit vector in the at least one unit vector according to the second distance information; determining a target unit vector based on the in-line point information corresponding to each unit vector; determining straight line segment information corresponding to the one point cluster according to the target unit vector to obtain at least one straight line segment information corresponding to the at least one point cluster; performing aggregation processing according to the at least one straight line segment information to obtain a road edge detection result.

2. The method of claim 1, wherein, The edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data comprises: performing preprocessing on the road surface point cloud data to obtain front view information corresponding to the road surface point cloud data; performing edge key point extraction processing according to the front view information to obtain the edge key point information.

3. The method of claim 2, wherein: the preprocessing is polar coordinate gridding processing or rearrangement processing; wherein the polar coordinate gridding processing is gridding processing according to polar coordinate information of points in the road surface point cloud data; and the rearrangement processing is processing of rearranging the road surface point cloud data into a row-column form.

4. The method of claim 1, wherein, The determination of the at least one point cluster according to the sorted key point information comprises: obtaining first distance information between each two adjacent edge key points in the sorted key point information; determining the at least one point cluster according to the first distance information and a first distance threshold.

5. The method of claim 1, wherein, The sorting processing on the edge key point information to obtain sorted key point information comprises: determining horizontal angle information corresponding to each edge key point in the edge key point information; performing sorting processing on each edge key point in the edge key point information according to the horizontal angle information to obtain the sorted key point information.

6. The method of claim 1, wherein, The in-line point information comprises in-line point quantity and straight line segment length information; wherein the straight line segment length information is length information of a straight line segment formed by the in-line points; and the determination of the target unit vector based on the in-line point information corresponding to each unit vector comprises: determining the target unit vector according to the in-line point quantity and the straight line segment length information.

7. The method of claim 1, wherein, The determination of in-line point information corresponding to each unit vector in the at least one unit vector according to the second distance information comprises: For one unit vector in the at least one unit vector, edge key points corresponding to distance information less than a second distance threshold in the second distance information corresponding to the one unit vector are taken as in-line points corresponding to the one unit vector, to obtain in-line point information corresponding to each unit vector in the at least one unit vector.

8. The method according to any one of claims 1 to 7, characterized in that, The aggregating processing according to the at least one straight line segment information obtains a road edge detection result, and the aggregating processing includes: determining target straight line segment information in the at least one straight line segment information, the target straight line segment information having a similarity greater than a similarity threshold; performing aggregating processing on the target straight line segment information to obtain the road edge detection result.

9. An electronic device, comprising: The electronic device includes an acquisition unit and a processing unit, The acquisition unit is configured to acquire road surface point cloud data. The processing unit is configured to perform edge key point extraction processing based on the road surface point cloud data to obtain edge key point information corresponding to the road surface point cloud data, perform sorting processing on the edge key point information to obtain sorted key point information, determine at least one point cluster according to the sorted key point information, determine one unit vector according to each two adjacent edge key points in the one point cluster to obtain at least one unit vector, calculate second distance information between each unit vector in the at least one unit vector and each edge key point in the one point cluster, determine in-line point information corresponding to each unit vector in the at least one unit vector according to the second distance information, determine target unit vector based on the in-line point information corresponding to each unit vector, determine straight line segment information corresponding to the one point cluster according to the target unit vector to obtain at least one straight line segment information corresponding to the at least one point cluster, and perform aggregating processing according to the at least one straight line segment information to obtain a road edge detection result. The electronic device includes a processor and a memory storing instructions executable by the processor; when the instructions are executed by the processor, the method in any one of claims 1-8 is implemented.

11. A computer readable storage medium having a program stored thereon, applied in an electronic device, and the program is executed by a processor to implement the method in any one of claims 1-8. ​ ​ ​ 10. An electronic device, comprising: ​ ​

Images