Road Edge Detection Method, Device, Equipment, Vehicle and Storage Medium

Abstract

The present application provides a curb detection method, device, equipment, vehicle and storage medium. The method includes: obtaining point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device arranged on the vehicle to detect the surrounding environment of the vehicle; performing polar coordinate rasterization on the point cloud to obtain a raster set; the raster in the raster set includes the center of gravity of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points; screening out a curb candidate raster set from the raster set according to the height difference; performing curb detection according to the relative distance relationship between the centers of gravity of the curb candidate rasters adjacent in angle in the curb candidate raster set. This embodiment realizes obtaining a curb result with continuity and smoothness.

Description

Technical Field

[0001] The present application relates to the technical field of point cloud processing, and particularly to a curb detection method, device, equipment, vehicle and storage medium. Background Art

[0002] In urban scenarios, curbs are one of the common scenario facilities, used to distinguish roads from sidewalks, roads from greenery, etc. During the autonomous driving process of a vehicle, accurately identifying curb information isolates the drivable area and non-drivable area for the autonomous driving vehicle, which is beneficial for the vehicle to perform path planning more safely and reliably, complete more complex functions and tasks, and also helps to avoid safety risks such as scraping against curbs. Therefore, it is necessary to propose a method that can accurately detect curbs. Summary of the Invention

[0003] In view of this, the present application provides a curb detection method, device, equipment, vehicle and storage medium.

[0004] Specifically, the present application is implemented through the following technical solutions:

[0005] According to the first aspect of the embodiments of the present application, a curb detection method is provided, including:

[0006] Obtain the point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device arranged on the vehicle to detect the surrounding environment of the vehicle;

[0007] Perform polar coordinate rasterization on the point cloud to obtain a raster set; the raster in the raster set includes the centroid of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points;

[0008] Screen out a set of candidate curb rasters from the raster set according to the height difference;

[0009] Perform curb detection according to the relative distance relationship between the centroids of the candidate curb rasters adjacent in angle in the set of candidate curb rasters.

[0010] Optionally, the performing curb detection according to the relative distance relationship between the centroids of the candidate curb rasters adjacent in angle in the set of candidate curb rasters includes:

[0011] After selecting a candidate curb raster as the initial seed point from the set of candidate curb rasters, repeat the following steps in a specified direction until there is no candidate curb raster adjacent in angle to the seed point: if the centroid distance between the candidate curb raster as the seed point and the candidate curb raster adjacent in angle is less than a preset distance, use the candidate curb raster adjacent in angle as the new seed point;

[0012] Generate a curb result based on all the curb candidate grids that serve as seed points.

[0013] Optionally, after selecting a curb candidate grid as an initial seed point from the set of curb candidate grids, it further includes:

[0014] Repeat the following steps in a specified direction until the number of executions meets a preset number: If the centroid distance between the curb candidate grid serving as the initial seed point and the curb candidate grid with the smallest angle difference at the current time is greater than or equal to a preset distance, select a new curb candidate grid with the smallest angle difference from the remaining unselected curb candidate grids as compared with the initial seed point.

[0015] Optionally, the detection device is installed on one side of the vehicle and is used to collect environmental information on one side of the road;

[0016] The curb candidate grid serving as the initial seed point is the curb candidate grid that is at one boundary of the polar coordinates and is the closest to the center of the polar coordinates, and the specified direction includes the direction from one boundary of the polar coordinates to the other boundary;

[0017] Or

[0018] The curb candidate grid serving as the initial seed point is the curb candidate grid that is at a non-boundary of the polar coordinates and is the closest to the center of the polar coordinates, and the specified direction includes two directions where the initial seed point points to the two boundaries of the polar coordinates respectively.

[0019] Optionally, the detection device is installed at the front of the vehicle and is used to collect environmental information in front of the road and on both sides of the road;

[0020] After screening out the set of curb candidate grids from the grid set according to the height difference, it further includes:

[0021] Divide the set of curb candidate grids into two subsets of curb candidate grids according to the dividing line extending from the center of the polar coordinates in the vehicle driving direction;

[0022] Among them, in each subset of curb candidate grids, the curb candidate grid serving as the initial seed point is the curb candidate grid that is at the boundary of the polar coordinates and is the closest to the center of the polar coordinates, and the specified direction includes the vehicle driving direction; or, the curb candidate grid serving as the initial seed point is the curb candidate grid that is at a non-boundary of the polar coordinates and is the closest to the center of the polar coordinates, and the specified direction includes the direction where the initial seed point points to the boundary of the polar coordinates and the vehicle driving direction;

[0023] The generating of the curb result based on all the curb candidate grids that serve as seed points includes:

[0024] For each of the curb candidate grid sub-sets, curb results are respectively generated according to all the curb candidate grids that serve as seed points in the curb candidate grid sub-set.

[0025] Optionally, after screening out the curb candidate grid sub-set from the grid set according to the height difference, the following is further included:

[0026] In the curb candidate grid sub-set, delete the remaining curb candidate grids except the curb candidate grid closest to the polar coordinate center at each polar coordinate angle.

[0027] Optionally, the following is further included:

[0028] For each curb candidate grid in the curb candidate grid sub-set, use the median or average value of the distances from the curb candidate grid and at least one curb candidate grid adjacent to it in terms of angle to the polar coordinate center as the distance from the curb candidate grid to the polar coordinate center.

[0029] Optionally, the screening out of the curb candidate grid sub-set from the grid set according to the height difference includes:

[0030] Determine the grids with the height difference greater than the second preset height as curb candidate grids; wherein, the second preset height indicates the preset height difference between the curb and the ground.

[0031] Optionally, the grid further includes the maximum height and the minimum height among all the three-dimensional points;

[0032] The screening out of the curb candidate grid sub-set from the grid set according to the height difference includes:

[0033] For each grid in the grid set, determine at least one adjacent grid adjacent to it in terms of distance to obtain a grid sub-set;

[0034] For each of the grid sub-sets, if the height difference between the maximum height and the minimum height in the grid sub-set is greater than the second preset height, determine curb candidate grids according to the grid sub-set; wherein, the second preset height indicates the preset height difference between the curb and the ground.

[0035] Optionally, before rasterizing the point cloud into a grid set in polar coordinates, the following is further included:

[0036] Process the point cloud using a dynamic object detection network, identify and remove the three-dimensional points indicating dynamic objects in the point cloud; and / or

[0037] Remove the three-dimensional points in the point cloud with a height greater than the first preset height, where the first preset height indicates a height higher than the curb.

[0038] According to a second aspect of the embodiments of the present application, a curb detection device is provided, including:

[0039] A point cloud acquisition module for acquiring a point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device disposed on a vehicle detecting the surrounding environment of the vehicle;

[0040] A rasterization module for performing polar coordinate rasterization on the point cloud to obtain a raster set; the raster in the raster set includes the centroid of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points;

[0041] A screening module for screening out a curb candidate raster set from the raster set according to the height difference;

[0042] A curb detection module for performing curb detection according to the relative distance relationship between the centroids of adjacent curb candidate rasters in the angle in the curb candidate raster set. According to a third aspect of the embodiments of the present application, an electronic device is provided, including a memory, a processor, and an executable instruction stored on the memory and executable on the processor;

[0043] Wherein, when the processor executes the executable instruction, the steps in the method according to any one of the first aspects are implemented.

[0044] According to a fourth aspect of the embodiments of the present application, a vehicle is provided, provided with the electronic device as described in the third aspect; or, the vehicle is independent of and communicatively connected to the electronic device as described in the third aspect.

[0045] According to a fifth aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer instruction is stored, and when the computer instruction is executed by a processor, the steps in the method according to any one of the first aspects are implemented.

[0046] The technical solutions provided by the embodiments of the present application may include the following beneficial effects:

[0047] The embodiments of the present application provide a curb detection method, device, equipment, vehicle and storage medium, which acquire a point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device disposed on a vehicle detecting the surrounding environment of the vehicle; after performing polar coordinate rasterization on the point cloud to obtain a raster set, considering that there is a certain height difference between the curb and the ground, therefore, a curb candidate raster set can be screened out from the raster set according to the height difference, which is beneficial to improving the accuracy of curb detection; and then considering the continuity characteristic of the curb, curb detection is performed according to the relative distance relationship between the centroids of adjacent curb candidate rasters in the angle in the curb candidate raster set, so that the determined curb detection result has smoothness and continuity.

[0048] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and should not limit this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application.

[0050] Figure 1 is a schematic structural diagram of a vehicle and an electronic device shown in an exemplary embodiment of this application.

[0051] Figure 2 is another schematic structural diagram of a vehicle and an electronic device shown in an exemplary embodiment of this application.

[0052] Figure 3 is a schematic flowchart of a curb detection method shown in an exemplary embodiment of this application.

[0053] Figure 4 is a schematic diagram of a grid after point cloud polar coordinate rasterization shown in an exemplary embodiment of this application.

[0054] Figure 5A 、 Figure 5B and Figure 5C are schematic diagrams of three forms of the existence of a curb in a grid shown in an exemplary embodiment of this application.

[0055] Figure 6 is a schematic diagram of a detection device installed on one side of a vehicle detecting one side of a road shown in an exemplary embodiment of this application.

[0056] Figure 7 is a schematic diagram of a detection device installed at the front of a vehicle detecting the front and both sides of a road shown in an exemplary embodiment of this application.

[0057] Figure 8 is a schematic diagram of clustering curb candidate grids in a curb candidate grid subset by using the seed growth method shown in an exemplary embodiment of this application.

[0058] Figure 9 is a schematic structural diagram of a curb detection device shown in an exemplary embodiment of this application.

[0059] Figure 10 is a schematic structural diagram of an electronic device shown in an exemplary embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0061] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0062] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0063] In response to the problems in the related art, the embodiment of the present application provides a curb detection method, which can obtain a point cloud collected by a detection device such as a laser radar or other depth camera installed on a vehicle, and use the point cloud to perform curb detection. After the point cloud is rasterized in polar coordinates to obtain a grid set, considering that there is a certain height difference between the curb and the ground, a curb candidate grid set can be screened from the grid set according to the height difference, which is conducive to improving the accuracy of curb detection; and considering the continuity characteristics of the curb, the curb detection is performed according to the relative distance relationship between the centroids of angularly adjacent curb candidate grids in the curb candidate grid set, so that the determined curb detection result has smoothness and continuity.

[0064] In some embodiments, the roadside detection method provided in the embodiments of the present application can be applied to electronic devices. Figure 1 , for example, the electronic device 10 may be a vehicle-mounted terminal installed on a vehicle 20. Figure 2 The electronic device 10 and the vehicle 20 are independent of each other, and the electronic device 10 is communicatively connected to the vehicle 20; for example, the electronic device may be a server or a cloud server communicatively connected to the vehicle, or the electronic device may be other computing devices such as a mobile terminal or a computer communicatively connected to the vehicle.

[0065] Exemplarily, the electronic device may include a memory and a processor. The curb detection method may be executable instructions stored in the memory, and when the processor executes the executable instructions, the curb detection method may be implemented.

[0066] Exemplarily, the vehicle is equipped with a detection device (such as a lidar, an RGBD camera, a binocular vision sensor, a millimeter-wave radar, etc.). During the driving process of the vehicle, the detection device detects the current environment in real time to obtain the point cloud of the current environment. Taking the lidar as an example, the lidar is used to emit a laser pulse sequence to the current environment, then receive the laser pulse sequence reflected from the target, and generate a point cloud according to the reflected laser pulse sequence. In one example, the lidar may determine the reception time of the reflected laser pulse sequence. For example, the reception time of the laser pulse sequence is determined by detecting the rising edge time and / or the falling edge time of the detection electrical signal pulse. Thus, the lidar can calculate the TOF (Time of Flight) by using the reception time information and the emission time of the laser pulse sequence, so as to determine the distance from the detected object to the lidar. The lidar belongs to a sensor that emits light independently, does not depend on the light source illumination, is less affected by ambient light interference, and can work normally even in a lightless enclosed environment, having wide applicability.

[0067] Exemplarily, the point cloud includes a number of three-dimensional points, and each three-dimensional point includes position information and attribute information. The position information may be, for example, three-dimensional coordinates in a three-dimensional coordinate system established with the lidar as the center. The attribute information includes but is not limited to reflection intensity, surface normal vector, and / or color information, etc.

[0068] Next, an exemplary description of the curb detection method provided in the embodiments of the present application will be given. Please refer to Figure 3 , Figure 3 which is a schematic flow diagram of a curb detection method. The method includes:

[0069] In step S101, the point cloud to be processed is obtained; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device arranged on the vehicle detecting the surrounding environment of the vehicle.

[0070] In step S102, the point cloud is polar coordinate rasterized to obtain a raster set; the raster in the raster set includes the centroid of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points.

[0071] In step S103, the curb candidate raster set is screened out from the raster set according to the height difference.

[0072] In step S104, curb detection is performed according to the relative distance relationship between the centroids of adjacent curb candidate grids in the curb candidate grid set. In this embodiment, after rasterizing the point cloud into a grid set in polar coordinates, considering that there is a certain height difference between the curb and the ground, the curb candidate grid set can be filtered out from the grid set according to the height difference, which is beneficial to improving the accuracy of curb detection; furthermore, considering the continuity characteristics of the curb, curb detection is performed according to the relative distance relationship between the centroids of adjacent curb candidate grids in the curb candidate grid set, so that the determined curb detection result has smoothness and continuity.

[0073] In some embodiments, after the electronic device obtains the point cloud collected by a detection device (such as a lidar) in the vehicle, it can first preprocess the point cloud to improve the accuracy of subsequent curb detection.

[0074] In a possible implementation manner, considering that the curb is a static object, the detection device will detect dynamic targets (such as other vehicles, pedestrians, bicycles, electric vehicles, etc.) in addition to static objects during the detection of the environment, and the dynamic targets are information irrelevant to curb detection. Therefore, in order to exclude the interference of the three-dimensional points indicating dynamic targets, the electronic device can use a preset dynamic target detection network to process the point cloud, identify the three-dimensional points indicating dynamic targets in the point cloud and remove them, which is beneficial to improving the accuracy of subsequent curb detection.

[0075] Exemplarily, the dynamic target detection network is used to extract features from the point cloud, so as to identify dynamic targets based on the extracted point cloud features, and then output a dynamic target recognition result, which includes three-dimensional points indicating dynamic targets, so that the electronic device can remove the three-dimensional points indicating dynamic targets in the point cloud based on the dynamic target recognition result. It can be understood that the structure of the dynamic target detection network in the embodiments of the present application is not limited in any way, and relevant detection networks or models with this function in related technologies can be referred to.

[0076] In another possible implementation manner, the detection device may detect objects higher than the curb (such as trees, traffic sign poles, etc.) in addition to the curb during the detection of the environment. The curb is located on the ground and is connected to the road in space, usually slightly higher than the ground but not too high. Therefore, in order to avoid the interference of three-dimensional points with too high height, the electronic device can remove the three-dimensional points in the point cloud with a height greater than a first preset height, and the first preset height indicates a height higher than the curb, so as to filter out a part of the height point cloud close to the ground, which is beneficial to improving the accuracy of subsequent curb detection.

[0077] It can be understood that the specific value of the first preset height can be specifically set according to the actual application scenario. For example, the common curb height is 5-30 cm. Therefore, the first preset height can be set to 35 cm, and the three-dimensional points with a height higher than 35 cm are removed, and the three-dimensional points with a set height below 35 cm are retained to participate in the curb detection.

[0078] After the electronic device obtains the point cloud collected by the detection device (such as lidar), it can perform at least one of the above preprocessing processes on the point cloud according to the actual situation.

[0079] In some embodiments, considering that the three-dimensional points in the point cloud are usually disordered, in order to facilitate the subsequent curb detection process, the electronic device can perform rasterization processing on the point cloud, so as to convert the disordered point cloud into an ordered grid. Taking lidar as an example, after lidar scans the environment, the point cloud is usually dense near and sparse far away; the inventor found that the polar coordinate system has a similar law to the scanning range of lidar. The polar coordinate system refers to taking a fixed point O in the plane as the pole, drawing a ray Ox as the polar axis, and then selecting a length unit and the positive direction of the angle (usually taking the counterclockwise direction). For any point M in the plane, use ρ to represent the length of the line segment OM, θ to represent the angle from Ox to OM, ρ is called the polar radius of point M, θ is called the polar angle of point M, and the ordered pair (ρ, θ) is called the polar coordinate of point M. The coordinate system established in this way is called the polar coordinate system. Please refer to Figure 4 , after dividing the polar coordinates according to the angle and distance, the area of the grid closer to the center of the polar coordinates is smaller, and the area of the grid farther from the center of the polar coordinates is larger. Therefore, in step S101 of the embodiment of the present application, the electronic device performs polar coordinate rasterization on the point cloud to obtain a grid set, so as to obtain more and more accurate point cloud information from the grid set and improve the accuracy of subsequent curb detection. In an example, please refer to Figure 4 , first define the angle resolution (angle_resolution) and range resolution (range_resolution) of the polar coordinate system; for example, the angle resolution is 0.1°, which means that each 0.1° is divided into an angle, and a sector is formed with the coordinate origin as the center; of course, the angle can also be divided according to other angle values (such as 1°, 2°, etc.). For example, the range resolution is 0.1 m, which means that within each angular sector, a grid is divided every 0.1 m, and each grid has a corresponding polar coordinate; of course, the distance can also be divided according to other distance values (such as 0.5 m, 0.2 m, etc.). For each three-dimensional point (x, y, z) in the point cloud, let the angle of the three-dimensional point projected onto the polar coordinate system be grid_x, and the distance of the three-dimensional point projected onto the polar coordinate system be grid_y, then the coordinates of the three-dimensional point in the polar coordinate system can be calculated as:

[0080]

[0081] After calculating the grid to which each three-dimensional point in the point cloud belongs, the electronic device may update the grid attributes. Each grid in the grid set includes the centroid of all three-dimensional points projected onto the grid, the height difference between the maximum height and the minimum height among all the three-dimensional points, and the set of all three-dimensional points projected onto the grid. Among them, the centroid of all three-dimensional points projected onto the grid is the average value of the three-dimensional coordinates of all the three-dimensional points. In one example, assuming the centroid is (X, Y, Z), and there are n three-dimensional points projected onto this grid, with three-dimensional coordinates (x1, y1, z1), (x2, y2, z2),..., (xn, yn, zn) respectively, then The maximum height refers to the maximum z value among all the three-dimensional points, and the minimum height refers to the minimum z value among all the three-dimensional points. Of course, in addition to including the above attribute information, other attribute information may also be included, such as the maximum height and the minimum height.

[0082] In some embodiments, after performing polar coordinate rasterization on the point cloud, in step S102, considering that the curb is higher than the ground, the electronic device can filter out the curb candidate grid set from the grid set according to the height difference, which is beneficial to the accuracy of the subsequent curb detection result.

[0083] Exemplarily, please refer to Figure 5A and Figure 5B . Considering the two situations of the curb in the grid, in Figure 5A , the complete jump process of the curb is visible in a single grid. In Figure 5B , the height difference within the grid is relatively small, so this grid will be classified as a ground attribute grid. In other words, the electronic device may determine the grid with a height difference greater than a second preset height as a curb candidate grid; where the second preset height indicates the preset height difference between the curb and the ground, and can be specifically set according to the actual ground construction situation. For example, the common curb height is 5 - 30 cm, so the second preset height can be set to 5 cm.

[0084] Exemplarily, please refer to Figure 5C, there is a third case where the curb is located at the edge of the grid. For this case, the grid also includes the maximum height and the minimum height among all the three-dimensional points; for each grid in the grid set, the electronic device determines at least one adjacent grid adjacent to the grid to obtain a sub-grid set; then for each of the sub-grid sets, if the height difference between the maximum height and the minimum height in the sub-grid set is greater than the second preset height, the curb candidate grid is determined according to the sub-grid set. In this embodiment, the height differences of at least two consecutive grids are comprehensively evaluated to ensure that the curb can be completely covered, avoid the missed detection phenomenon caused by the curb appearing at the grid boundary, and improve the accuracy of grid detection.

[0085] In one example, the electronic device may determine all the grids in the sub-grid set as curb candidate grids. In another example, considering that the grids closer to the polar coordinate center in the sub-grid set have a higher probability of belonging to the ground attribute grids, in order to save the subsequent operation data volume, the electronic device may determine the grids far from the coordinate center in the sub-grid set as curb candidate grids, that is, exclude the grids close to the coordinate center in the sub-grid set.

[0086] In one example, in the polar coordinate grid, for each angle from near to far, except for the first grid and the last grid of the angle, for each grid itself i, the adjacent grid i-1 close to the coordinate center, and the adjacent grid i+1 far from the coordinate center, the electronic device determines the height difference between the maximum height and the minimum height among these three consecutive grids. If the height difference is greater than the second preset height, in order to save the subsequent operation data volume, the grid i and / or the adjacent grid i+1 far from the coordinate center are determined as curb candidate grids.

[0087] In some embodiments, after the curb candidate grid set is obtained by screening based on the above method, considering that among the curb candidate grids at the same angle in the polar coordinates, the curb candidate grids far from the polar coordinate center will be blocked by the curb candidate grids close to the polar coordinate center, the electronic device preferentially considers the curb candidate grids close to the polar coordinate center. Therefore, in order to save the subsequent operation data volume, in the curb candidate grid set, the electronic device retains the curb candidate grid closest to the polar coordinate center at each polar coordinate angle, and then deletes the other curb candidate grids except the curb candidate grid closest to the polar coordinate center at each polar coordinate angle.

[0088] In some embodiments, the curb candidate grid set includes two types of grids: grids correctly classified as curbs and grids misclassified as curbs. Therefore, in order to reduce the influence of misclassified grids, for each curb candidate grid in the curb candidate grid set, the electronic device uses the median or average of the distances from the curb candidate grid and at least one curb candidate grid adjacent to it in terms of angle to the polar coordinate center as the distance from the curb candidate grid to the polar coordinate center. In this embodiment, median filtering or mean filtering is used to effectively filter the noise and burrs of the curb grids.

[0089] In one example, for each curb candidate grid i in the curb candidate grid set, the electronic device examines at least 2 curb candidate grids adjacent to the left and right angles of the curb candidate grid (such as curb candidate grid i - 1 and curb candidate grid i + 1), and performs median filtering or mean filtering on the distances from the grids to the center, that is, takes the median or average of the distances from these at least 3 curb candidate grids (i - 1, i, i + 1) to the center, and sets it as the distance from the curb candidate grid i to the center. After performing the above median filtering or mean filtering on each curb candidate grid in the curb candidate grid set, the burrs and noise at consecutive angles are effectively filtered.

[0090] In some embodiments, in step S104, the electronic device performs curb detection based on the relative distance relationship between the centroids of the curb candidate grids adjacent in terms of angle in the curb candidate grid set.

[0091] In a possible implementation manner, if a continuous plurality of curb candidate grids all meet the following condition: the relative distance between the centroids of the curb candidate grids adjacent in terms of angle is less than a preset threshold, then a curb result is generated based on the continuous plurality of curb candidate grids.

[0092] In another possible implementation, after the electronic device selects a curb candidate grid from the set of curb candidate grids as an initial seed point, it repeatedly executes the following steps in a specified direction until there are no curb candidate grids that are angularly adjacent to the seed point: If the centroid distance between the curb candidate grid as the seed point and the angularly adjacent curb candidate grid is less than a preset distance, the angularly adjacent curb candidate grid is used as a new seed point. After executing the above steps, a curb result is generated based on all the curb candidate grids that serve as seed points. This embodiment takes into account the continuity characteristics of the curb and uses the seed growth method to cluster the curb candidate grids in the set of curb candidate grids. Whether the centroid distance between the curb candidate grid as the seed point and the angularly adjacent curb candidate grid meets the requirements is used to determine a new seed point for continued growth, thereby further screening out the grids where there are curbs; finally, a curb result is generated based on all the curb candidate grids that serve as seed points. For example, all the curb candidate grids that serve as seed points can be connected in sequence to obtain the curb result, which can ensure the smoothness and continuity of the generated curb result.

[0093] After the electronic device selects a curb candidate grid from the set of curb candidate grids as an initial seed point, it determines the curb candidate grids that are angularly adjacent to the curb candidate grid as the seed point in a specified direction. If the centroid distance between the curb candidate grid as the seed point and the angularly adjacent curb candidate grid is less than a preset distance, the angularly adjacent curb candidate grid is used as a new seed point for continued growth; then it continues to determine the curb candidate grids that are angularly adjacent to the new seed point in the specified direction, compares the centroid distance between the new seed point and the angularly adjacent curb candidate grid with the preset distance. If it is less than the preset distance, the angularly adjacent curb candidate grid is used as a new seed point for continued growth, and so on, until there are no curb candidate grids that are angularly adjacent to the seed point in the specified direction. This embodiment uses the seed growth method to cluster the curb candidate grids in the set of curb candidate grids and determines the continuous relationship between the seed points through the centroid distance, so that a continuous and smooth curb result can be generated based on the selected seed points.

[0094] For example, assume that the curb candidate grid as the initial seed point is i. When growing along the specified direction, if the centroid distance between the curb candidate grid i and the curb candidate grid i+1 adjacent in terms of angle is less than the preset distance, then the curb candidate grid i+1 is taken as the new seed point to continue growing; if the centroid distance between the curb candidate grid i+1 and the curb candidate grid i+2 adjacent in terms of angle is less than the preset distance, then the curb candidate grid i+2 is taken as the new seed point to continue growing; if the centroid distance between the curb candidate grid i+2 and the curb candidate grid i+3 adjacent in terms of angle is less than the preset distance, then the curb candidate grid i+3 is taken as the new seed point to continue growing; and so on, until there is no curb candidate grid i+n+1 adjacent in terms of angle to the curb candidate grid i+n as the seed point along the specified direction, where n is an integer greater than 0.

[0095] After the electronic device selects the curb candidate grid as the initial seed point from the set of curb candidate grids, it determines the curb candidate grid with the smallest angular difference from the curb candidate grid as the initial seed point along the specified direction. The first curb candidate grid with the smallest angular difference from the initial seed point is the curb candidate grid adjacent in terms of angle to the initial seed point. If the centroid distance between the curb candidate grid as the initial seed point and the curb candidate grid adjacent in terms of angle is greater than or equal to the preset distance, the electronic device selects a new curb candidate grid with the smallest angular difference from the initial seed point from the remaining unselected curb candidate grids, and then compares the centroid distance between the initial seed point and the current curb candidate grid with the smallest angular difference with the preset distance. If it is still greater than or equal to the preset distance, the electronic device continues to select a new curb candidate grid with the smallest angular difference from the initial seed point from the remaining unselected curb candidate grids for inspection, and so on, until the number of executions meets the preset number, then the curb stops growing. In this embodiment, the seed points are screened by the centroid distance, which is beneficial to selecting seed points with a continuous relationship and excluding curb candidate grids without a continuous relationship.

[0096] For example, assume that the curb candidate grid as the initial seed point is i, and it grows along the specified direction. If the centroid distance between the curb candidate grid i and the adjacent curb candidate grid i+1 at an angle is greater than or equal to the preset distance, a new curb candidate grid i+2 with the smallest angle difference from the initial seed point is selected from the remaining unselected curb candidate grids; if the centroid distance between the curb candidate grid i and the curb candidate grid i+2 is greater than or equal to the preset distance, a new curb candidate grid i+3 with the smallest angle difference from the initial seed point is selected from the remaining unselected curb candidate grids; if the centroid distance between the curb candidate grid i and the curb candidate grid i+3 is greater than or equal to the preset distance, a new curb candidate grid i+4 with the smallest angle difference from the initial seed point is selected from the remaining unselected curb candidate grids for further investigation; and so on. If no curb candidate grid that meets the preset distance is found after examining i+m (or performing m times), the growth of the curb stops; m is specifically set according to the actual application scenario.

[0097] In an exemplary embodiment, please refer to Figure 6 , the detection device 21 for collecting point clouds is installed on one side of the vehicle 20 and can collect the environmental information on one side of the road. Then, the electronic device can determine a curb result from the set of curb candidate grids.

[0098] Exemplarily, in the process of clustering the curb candidate grids in the set of curb candidate grids by using the seed growth method, the selected curb candidate grid as the initial seed point can be the curb candidate grid at one boundary of the polar coordinate and closest to the center of the polar coordinate. Further, the initial seed point can grow in the specified direction, and the specified direction can be the direction from one boundary of the polar coordinate to the other boundary. Herein, one boundary and the other boundary of the polar coordinate refer to the angular boundaries, such as Figure 4 the straight boundaries on the left and right sides in. The curb candidate grids on the two boundaries of the polar coordinate can indicate the curb information detected by the two detection boundaries of the detection device. Exemplarily, in the process of clustering the curb candidate grids in the set of curb candidate grids by using the seed growth method, the selected curb candidate grid as the initial seed point can be the curb candidate grid at a non-boundary of the polar coordinate and closest to the center of the polar coordinate. Further, the initial seed point can grow in the specified direction, and the specified direction includes two directions corresponding to the initial seed point pointing to the two boundaries of the polar coordinate respectively. That is to say, the initial seed point grows towards the two boundaries of the polar coordinate respectively.

[0099] After determining all the seed points, the electronic device can connect all the curb candidate grids serving as seed points in sequence to obtain the curb result.

[0100] In another exemplary embodiment, please refer toFigure 7 The detection device 21 for collecting point clouds is installed at the front of the vehicle 20 and can collect environmental information in front of and on both sides of the road. Then, the electronic device may be able to determine two road edge results from the set of road edge candidate grids, which are located on the left and right sides of the vehicle respectively. Before clustering using the seed growth method, considering that road edges are usually located on the left and right sides of the vehicle, the electronic device first divides the set of road edge candidate grids into two subsets of road edge candidate grids according to the dividing line extending from the polar coordinate center in the vehicle driving direction, and then uses the seed growth method to cluster the road edge candidate grids in the two subsets of road edge candidate grids respectively. The clustering method is similar to the above-described process and will not be elaborated here, so that two road edge results located on the left and right sides of the vehicle can be determined.

[0101] Exemplarily, during the process of clustering the road edge candidate grids in the subset of road edge candidate grids using the seed growth method, the road edge candidate grid selected as the initial seed point may be the road edge candidate grid at the polar coordinate boundary and closest to the polar coordinate center. Furthermore, the initial seed point can grow in a specified direction, and the specified direction may be the vehicle driving direction. Please refer to Figure 8 , Figure 8 shows the road edge result determined by the initial seed point at the polar coordinate boundary growing in the vehicle driving direction.

[0102] Exemplarily, during the process of clustering the road edge candidate grids in the subset of road edge candidate grids using the seed growth method, the road edge candidate grid selected as the initial seed point may also be the road edge candidate grid at the non-polar coordinate boundary and closest to the polar coordinate center. Furthermore, the initial seed point can grow in a specified direction, and the specified direction includes the direction from the initial seed point to the polar coordinate boundary and the vehicle driving direction, that is, the initial seed point grows in both directions towards the polar coordinate boundary and the vehicle driving direction.

[0103] For each of the subsets of road edge candidate grids, after determining all the seed points, the electronic device generates a road edge result according to all the road edge candidate grids serving as seed points in this subset of road edge candidate grids; for example, connecting all the road edge candidate grids serving as seed points in sequence to obtain the road edge result.

[0104] It is not difficult to understand that the solutions described in the above embodiments can be combined without conflict, and they are not listed one by one in the embodiments of the present application.

[0105] Correspondingly, please refer to Figure 9 The embodiments of the present application also provide a road edge detection device, including:

[0106] The point cloud acquisition module 201 is configured to acquire a point cloud to be processed; the point cloud includes a plurality of three-dimensional points, and the point cloud is obtained by a detection device disposed on a vehicle detecting the surrounding environment of the vehicle.

[0107] The rasterization module 202 is configured to perform polar coordinate rasterization on the point cloud to obtain a raster set; the raster in the raster set includes the centroid of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points.

[0108] The screening module 203 is configured to screen a curb candidate raster set from the raster set according to the height difference.

[0109] The curb detection module 204 is configured to perform curb detection according to the relative distance relationship between the centroids of adjacent curb candidate rasters in the curb candidate raster set in terms of angle.

[0110] In some embodiments, the curb detection module 204 includes a seed growth unit and a curb result generation unit.

[0111] The seed growth unit is configured to, after selecting a curb candidate raster as an initial seed point from the curb candidate raster set, repeatedly perform the following steps along a specified direction until there is no curb candidate raster adjacent to the seed point in terms of angle: if the centroid distance between the curb candidate raster as the seed point and the adjacent curb candidate raster in terms of angle is less than a preset distance, use the adjacent curb candidate raster in terms of angle as a new seed point.

[0112] The curb result generation unit is configured to generate a curb result according to all the curb candidate rasters as seed points.

[0113] In some embodiments, the seed growth unit is further configured to repeatedly perform the following steps along a specified direction until the number of executions meets a preset number: if the centroid distance between the curb candidate raster as the initial seed point and the curb candidate raster with the smallest current angle difference is greater than or equal to the preset distance, select a new curb candidate raster with the smallest angle difference from the remaining unselected curb candidate rasters with respect to the initial seed point.

[0114] In some embodiments, the detection device is installed on one side of the vehicle and is configured to collect environmental information on one side of the road; the curb candidate raster as the initial seed point is the curb candidate raster at one boundary of the polar coordinate and closest to the center of the polar coordinate, and the specified direction includes the direction from one boundary of the polar coordinate to the other boundary; or the curb candidate raster as the initial seed point is the curb candidate raster at a non-boundary of the polar coordinate and closest to the center of the polar coordinate, and the specified direction includes two directions from the initial seed point to the two corresponding boundaries of the polar coordinate.

[0115] In some embodiments, the detection device is installed at the front of the vehicle and is used to collect environmental information in front of and on both sides of the road. The device further includes a division module, which is used to divide the set of curb candidate grids into two subsets of curb candidate grids according to a demarcation line extending from the polar coordinate center in the vehicle traveling direction.

[0116] Wherein, in each subset of curb candidate grids, the curb candidate grid serving as the initial seed point is the curb candidate grid that is at the polar coordinate boundary and closest to the polar coordinate center, and the specified direction includes the vehicle traveling direction; alternatively, the curb candidate grid serving as the initial seed point is the curb candidate grid that is at the non-polar coordinate boundary and closest to the polar coordinate center, and the specified direction includes the direction from the initial seed point to the polar coordinate boundary and the vehicle traveling direction.

[0117] The curb result generation unit is specifically configured to, for each subset of curb candidate grids, generate a curb result according to all the curb candidate grids serving as seed points in the subset of curb candidate grids.

[0118] In some embodiments, the device further includes a deletion module, which is used to delete, in the set of curb candidate grids, the remaining curb candidate grids except the curb candidate grid closest to the polar coordinate center at each polar coordinate angle.

[0119] In some embodiments, the device further includes a filtering module, which is used to, for each curb candidate grid in the set of curb candidate grids, take the median or average value of the distances from the curb candidate grid and at least one curb candidate grid adjacent to it in terms of angle to the polar coordinate center as the distance from the curb candidate grid to the polar coordinate center.

[0120] In some embodiments, the screening module 203 is specifically configured to determine a grid with a height difference greater than a second preset height as a curb candidate grid; wherein, the second preset height indicates a preset height difference between the curb and the ground.

[0121] In some embodiments, the grid further includes the maximum height and the minimum height among all the three-dimensional points. The screening module 203 is specifically configured to, for each grid in the grid set, determine at least one adjacent grid adjacent to the grid in terms of distance to obtain a subset of grids; for each subset of grids, if the height difference between the maximum height and the minimum height in the subset of grids is greater than the second preset height, determine a curb candidate grid according to the subset of grids; wherein, the second preset height indicates a preset height difference between the curb and the ground.

[0122] In some embodiments, the device further includes a point cloud preprocessing module, which is configured to process the point cloud using a dynamic target detection network, identify three-dimensional points in the point cloud indicating dynamic targets and remove them; and / or remove three-dimensional points in the point cloud with a height greater than a first preset height, where the first preset height indicates a height higher than the curb.

[0123] For the implementation processes of the functions and roles of each unit in the above device, please refer to the implementation processes of the corresponding steps in the above method for details, which will not be elaborated here.

[0124] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial descriptions of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0125] Correspondingly, please refer to Figure 10 , this application embodiment also provides an electronic device 10, including a memory 11, a processor 12, and executable instructions stored on the memory 11 and executable on the processor 12;

[0126] Wherein, when the processor 12 executes the executable instructions, the steps in any one of the above methods are implemented.

[0127] Exemplarily, the processor 12 includes, but is not limited to, a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), or a field-programmable gate array (FPGA), etc.

[0128] Exemplarily, the memory 11 may include at least one type of storage medium, and the storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc.

[0129] For the implementation processes of the functions and roles of the various components in the above device, please refer to the implementation processes of the corresponding steps in the above method for details, which will not be elaborated here.

[0130] Correspondingly, please refer to Figure 1 or Figure 2 , an embodiment of the present application further provides a vehicle, which is provided with the above electronic device; or, the vehicle is independent of and communicatively connected to the above electronic device.

[0131] It can be understood that the vehicle further includes other components. Generally, a vehicle includes a chassis, a body, an engine, and electrical equipment. The engine is the power device of the vehicle and is used to generate power; the chassis is used to support the engine and the body, and the chassis can drive the vehicle to move according to the power generated by the engine; the body is installed on the frame of the chassis for the driver and passengers to ride or load goods; the electrical equipment includes a power source and electrical appliances. For example, the power source includes a storage battery and a generator, and the electrical appliances include the starting system of the engine or other electrical devices. Optionally, the vehicle further includes on-vehicle sensors (such as cameras, lidar, millimeter-wave radars, RGBD cameras, etc.) for sensing environmental information around the vehicle. Optionally, the vehicle further includes an autonomous driving system for assisting the driver in driving.

[0132] Correspondingly, an embodiment of the present application further provides a computer program product, including a computer program, which is used to implement the above curb detection method when executed by a processor.

[0133] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided. For example, a memory including instructions, and the above instructions can be executed by a processor of the device to complete the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0134] A non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by a processor of a terminal, enables the terminal to execute the above method.

[0135] The embodiments of the subject matter and the functional operations described in this specification can be implemented in digital electronic circuitry, tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or one or more combinations of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible non-transitory program carrier to be executed by, or to control the operation of, a data processing apparatus. Alternatively or additionally, the program instructions can be encoded on an artificially generated propagated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal, that is generated to encode and transmit information to the appropriate receiver apparatus for execution by the data processing apparatus. A computer storage medium may be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or one or more combinations of them.

[0136] The processes and logical flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform the functions by operating on input data and generating output. The processes and logical flows can also be performed by, or the apparatus can be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

[0137] Suitable computers for executing computer programs include, by way of example, general and / or special purpose microprocessors, or any other type of central processing unit. Generally, a central processing unit will receive instructions and data from a read only memory and / or a random access memory. Basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, etc., or the computer will be operatively coupled to such mass storage devices to receive data from them or to transfer data to them, or both. However, a computer need not have such devices. In addition, a computer may be embedded in another device, such as a mobile telephone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name just a few.

[0138] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, including, for example, semiconductor memory devices (such as EPROM, EEPROM, and flash memory devices), magnetic disks (such as internal hard disks or removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. Processors and memories can be supplemented by, or incorporated in, special purpose logic circuitry.

[0139] Although this specification contains many specific implementation details, these should not be construed as limiting the scope of any invention or the scope of what is claimed, but rather as mainly describing the features of specific embodiments of particular inventions. Certain features that are described in multiple embodiments in this specification can also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment can also be implemented separately in multiple embodiments or in any suitable sub-combination. In addition, although features may operate in certain combinations as described above and even be claimed as such initially, one or more features from the claimed combination can in some cases be removed from the combination, and the claimed combination can be directed to a sub-combination or a variation of a sub-combination.

[0140] Similarly, although operations are depicted in the drawings in a particular order, this should not be understood as requiring that the operations be performed in the particular order shown or sequentially, or that all illustrated operations be performed, to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of the various system modules and components in the above embodiments should not be understood as required in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0141] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the acts recited in the claims can be performed in a different order and still achieve the desired result. In addition, the processes depicted in the drawings are not necessarily in the particular order or sequential order shown to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0142] The above is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the scope of protection of the present application.

Claims

1. A curb detection method, characterized in that, Including: Obtaining point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device disposed on a vehicle detecting the environment around the vehicle; Performing polar coordinate rasterization on the point cloud to obtain a raster set; the raster in the raster set includes the centroid of all three-dimensional points projected onto the raster, and the height difference between the maximum height and the minimum height among all the three-dimensional points; Screening out a curb candidate raster set from the raster set according to the height difference; Performing curb detection according to the relative distance relationship between the centroids of adjacent curb candidate rasters in the curb candidate raster set in terms of angle; wherein, if the relative distance between the centroids of adjacent curb candidate rasters in terms of angle is less than a preset distance, there is a curb for the adjacent curb candidate rasters in terms of angle.

2. The method according to claim 1, characterized in that, The performing curb detection according to the relative distance relationship between the centroids of adjacent curb candidate rasters in the curb candidate raster set includes: After selecting a curb candidate raster as an initial seed point from the curb candidate raster set, repeating the following steps in a specified direction until there is no curb candidate raster adjacent to the seed point in terms of angle: if the centroid distance between the curb candidate raster as the seed point and the adjacent curb candidate raster in terms of angle is less than the preset distance, taking the adjacent curb candidate raster in terms of angle as a new seed point; Generating a curb result according to all the curb candidate rasters as seed points.

3. The method according to claim 2, wherein After selecting a curb candidate raster as an initial seed point from the curb candidate raster set, it further includes: Repeating the following steps in a specified direction until the number of executions meets a preset number: if the centroid distance between the curb candidate raster as the initial seed point and the curb candidate raster with the smallest current angle difference is greater than or equal to the preset distance, selecting a new curb candidate raster with the smallest angle difference from the remaining unselected curb candidate rasters with respect to the initial seed point.

4. The method according to claim 2 or 3, characterized in that, The detection device is installed on one side of the vehicle and is used for collecting environmental information on one side of the road; The curb candidate raster as the initial seed point is the curb candidate raster at one boundary of the polar coordinate and closest to the center of the polar coordinate, and the specified direction includes the direction from one boundary of the polar coordinate to the other boundary; Or The curb candidate raster as the initial seed point is the curb candidate raster at a non-boundary of the polar coordinate and closest to the center of the polar coordinate, and the specified direction includes two directions from the initial seed point to the two boundaries corresponding to the polar coordinate respectively.

5. The method according to claim 2 or 3, characterized in that, The detection device is installed at the front of the vehicle and is used for collecting environmental information in front of the road and on both sides of the road; After screening out the curb candidate raster set from the raster set according to the height difference, it further includes: Dividing the curb candidate raster set into two curb candidate raster subsets according to the demarcation line extending from the center of the polar coordinate in the vehicle driving direction; Among each of the curb candidate grid sub-sets, the curb candidate grid serving as the initial seed point is the curb candidate grid that is at the polar coordinate boundary and closest to the polar coordinate center, and the specified direction includes the vehicle driving direction; or, the curb candidate grid serving as the initial seed point is the curb candidate grid that is not at the polar coordinate boundary and closest to the polar coordinate center, and the specified direction includes the direction from the initial seed point to the polar coordinate boundary and the vehicle driving direction; The generating the curb result according to all the curb candidate grids serving as seed points includes: For each of the curb candidate grid sub-sets, generating a curb result according to all the curb candidate grids serving as seed points in this curb candidate grid sub-set respectively.

6. The method according to claim 1, characterized in that After screening out the curb candidate grid sub-set from the grid set according to the height difference, it further includes: In the curb candidate grid sub-set, deleting the remaining curb candidate grids except the curb candidate grid closest to the polar coordinate center at each polar coordinate angle.

7. The method according to claim 1 or 6, characterized in that, It further includes: For each curb candidate grid in the curb candidate grid sub-set, taking the median or average value of the distances from the curb candidate grid and at least one curb candidate grid adjacent to it in terms of angle to the polar coordinate center as the distance from the curb candidate grid to the polar coordinate center.

8. The method according to claim 1, wherein The screening out the curb candidate grid sub-set from the grid set according to the height difference includes: Determining the grids with the height difference greater than the second preset height as curb candidate grids; wherein, the second preset height indicates the preset height difference between the curb and the ground.

9. The method according to claim 1 or 8, characterized in that, The grid further includes the maximum height and the minimum height among all the three-dimensional points; The screening out the curb candidate grid sub-set from the grid set according to the height difference includes: For each grid in the grid set, determining at least one adjacent grid adjacent to this grid in terms of distance to obtain a grid sub-set; For each of the grid sub-sets, if the height difference between the maximum height and the minimum height in the grid sub-set is greater than the second preset height, determining the curb candidate grids according to the grid sub-set; wherein, the second preset height indicates the preset height difference between the curb and the ground.

10. The method according to claim 1, wherein Before rasterizing the point cloud into a grid set in polar coordinates, it further includes: Processing the point cloud using a dynamic object detection network to identify and remove the three-dimensional points indicating dynamic objects in the point cloud; and / or Removing the three-dimensional points in the point cloud with a height greater than the first preset height, where the first preset height indicates the height above the curb.

11. A curb detection device, characterized in that, It includes: A point cloud acquisition module for acquiring the point cloud to be processed; the point cloud includes a number of three-dimensional points, and the point cloud is obtained by a detection device arranged on the vehicle to detect the surrounding environment of the vehicle; A rasterization module for rasterizing the point cloud into a grid set in polar coordinates; the grid in the grid set includes the centroid of all the three-dimensional points projected onto this grid and the height difference between the maximum height and the minimum height among all the three-dimensional points; A screening module for screening out the curb candidate grid sub-set from the grid set according to the height difference; A curb detection module, configured to perform curb detection according to the relative distance relationship between the centroids of adjacent curb candidate grids in the curb candidate grid set; wherein, if the relative distance between the centroids of adjacent curb candidate grids is less than a preset distance, there is a curb between the adjacent curb candidate grids.

12. An electronic device, characterized in that, It includes a memory, a processor, and executable instructions stored on the memory and executable on the processor; Wherein, when the processor executes the executable instructions, the steps in the method according to any one of claims 1 to 10 are implemented.

13. A vehicle, characterized in that, An electronic device as claimed in claim 12 is provided; or, the vehicle and the electronic device as claimed in claim 12 are independent of each other and communicatively connected.

14. A computer-readable storage medium having computer instructions stored thereon, characterized in that, When the computer instructions are executed by the processor, the steps of the method according to any one of claims 1 to 10 are implemented.

Images