Methods, devices, storage media and electronic equipment for acquiring curb data

Abstract

This application relates to a method, apparatus, storage medium, and electronic device for acquiring roadside data. The method includes: acquiring target point cloud data of a mobile vehicle, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile vehicle within a target frame; dividing the target point cloud data into a grid network of the preset area; filtering out ground laser point clouds from each grid cell in the grid network to obtain the remaining laser point cloud; classifying the grid cells containing the remaining laser point cloud according to gradient information to obtain roadside data of the target frame; and merging the roadside data of the target frame with roadside data from historical frames to obtain the roadside data of the road where the mobile vehicle is located. This application solves the technical problem of being unable to acquire real-time and accurate roadside data.

Description

Technical Field

[0001] This application relates to the field of autonomous driving, and more particularly to a method, apparatus, storage medium, and electronic device for acquiring roadside data. Background Technology

[0002] With the maturity of autonomous driving technology, it has been widely applied in various industries. Driverless road sweepers are also increasingly being used in urban roads and park roads. These sweepers need to precisely follow the curb during operation. While traditional positioning systems can achieve edge-to-edge cleaning, interference from the Global Positioning System (GPS) signal or positioning errors can lead to miscalculations of the distance between the vehicle and the curb, resulting in the vehicle being too far or too close. Therefore, obtaining real-time and accurate curb data is crucial during automated cleaning. Summary of the Invention

[0003] This application provides a method, apparatus, storage medium, and electronic device for acquiring roadside data to solve the technical problem of being unable to acquire real-time and accurate roadside data.

[0004] In a first aspect, this application provides a method for acquiring roadside data, comprising: acquiring target point cloud data of a mobile vehicle, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile vehicle within a target frame; dividing the target point cloud data into a grid network of the preset area; filtering out ground laser point clouds in each grid cell of the grid network to obtain the remaining laser point cloud; classifying the grid cells including the remaining laser point cloud according to gradient information to obtain roadside data of the target frame; and merging the roadside data of the target frame with the roadside data of historical frames to obtain the roadside data of the road where the mobile vehicle is located.

[0005] To avoid interference from noisy point cloud data in the target point cloud data, before filtering out the ground laser point cloud in each grid of the grid network to obtain the remaining laser point cloud, obstacles on the road where the mobile vehicle is located can be determined; the target position of the obstacle within the target frame can be determined; and the grid corresponding to the target position in the grid network can be masked.

[0006] This embodiment identifies obstacles near the curb and removes the corresponding grid cells in the grid network, thereby removing obstacle interference during the curb data acquisition process and extracting accurate curb data.

[0007] If the target point cloud data is to be divided into a grid network of a preset area, one possible method is to divide the preset area into multiple blocks according to the plane where the road surface is located. The size of each block varies depending on its distance from the moving vehicle. The block in which each target point cloud data is located is then determined, thus completing the division of the target point cloud data.

[0008] One approach is to use a point on the mobile vehicle as the origin, and the direction of the vehicle's front and the direction perpendicular to the front as the horizontal and vertical coordinate axes; along one direction of the horizontal and vertical coordinate axes, the preset area is divided into equal sub-regions; along the other direction of the horizontal and vertical coordinate axes, each sub-region is divided into multiple blocks.

[0009] This example demonstrates how dividing the grid network by varying the size of the blocks based on their distance from the moving vehicle allows for the creation of a more detailed grid network for areas closer to the moving vehicle, thereby improving the accuracy of acquiring roadside data near the moving vehicle.

[0010] If we want to filter out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud, we can perform the following operations on each grid cell: fit the laser point cloud with a height difference less than a preset value into a plane; filter out the laser point cloud between the plane and the preset distance above the plane to obtain the remaining laser point cloud.

[0011] This example fits the laser point cloud within the grid with a height difference less than a preset value to a plane; filters out the laser point cloud between the plane and a preset distance above the plane to obtain the remaining laser point cloud. This allows for accurate determination of the ground laser point cloud within the grid, enabling its filtering and retention of accurate roadside laser point clouds.

[0012] If the grid including the remaining laser point cloud is classified according to the gradient information to obtain the path edge data of the target frame, the grid including the remaining laser point cloud can be traversed to calculate the vertical gradient of the grid; the grids with vertical gradients greater than the preset gradient and adjacent grids are merged into one connected grid; the merged one or more connected grids are determined as the path edge data of the target frame.

[0013] This embodiment merges scattered roadside data by merging the grid according to the vertical gradient, thereby obtaining complete roadside data.

[0014] If the road edge data of the target frame is merged with the road edge data of the historical frame to obtain the road edge data of the road where the mobile tool is located, the following method can be used: when the number of times a certain area is determined to be a road edge in the road edge data of the target frame and the road edge data of the historical frame exceeds a preset number, a certain area is determined to be a road edge; after determining whether each area is a road edge, the road edge data of the road is obtained.

[0015] This example obtains complete and accurate roadside data by comprehensively judging the roadside data of all frames, which can be used for the movement or operation of mobile tools.

[0016] Secondly, this application provides a roadside data acquisition device, comprising: an acquisition module for acquiring target point cloud data of a mobile vehicle, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile vehicle within a target frame; a segmentation module for segmenting the target point cloud data into a grid network of the preset area; a processing module for filtering out ground laser point clouds from each grid in the grid network to obtain remaining laser point clouds; a classification module for classifying the grids including the remaining laser point clouds according to gradient information to obtain roadside data of the target frame; and a merging module for merging the roadside data of the target frame with the roadside data of historical frames to obtain roadside data of the road where the mobile vehicle is located.

[0017] Thirdly, this application provides an electronic device, comprising: at least one communication interface; at least one bus connected to the at least one communication interface; at least one processor connected to the at least one bus; and at least one memory connected to the at least one bus, wherein the memory stores a computer program, and the processor is configured to implement the curb data acquisition method of any of the above when executing the computer program.

[0018] Fourthly, this application also provides a computer storage medium storing computer-executable instructions for performing the curb data acquisition method of any of the above claims of this application.

[0019] Fifthly, this application provides a mobile tool, including the aforementioned electronic device.

[0020] Compared with the prior art, the above-mentioned technical solution provided in this application has the following advantages: The solution provided in this application acquires the target point cloud data of the mobile tool, wherein the target point cloud data is the laser point cloud data of a preset area in front of the mobile tool within the target frame; divides the target point cloud data into a grid network of the preset area; filters out the ground laser point cloud in each grid in the grid network to obtain the remaining laser point cloud; classifies the grids including the remaining laser point cloud according to gradient information to obtain the road edge data of the target frame; merges the road edge data of the target frame with the road edge data of historical frames to obtain the road edge data of the road where the mobile tool is located, thereby extracting accurate road edge data based on the laser point cloud data in the preset area in front of the mobile tool, solving the technical problem in the prior art that it is impossible to obtain real-time and accurate road edge data. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0024] Figure 1 A flowchart illustrating a method for acquiring curb data provided in an embodiment of this application; Figure 2 A schematic diagram of curb data merging for a curb data acquisition method provided in this application embodiment; Figure 3 A flowchart illustrating another method for acquiring curb data provided in this application embodiment; Figure 4 A grid network diagram illustrating a method for acquiring curb data provided in this application embodiment; Figure 5 A schematic diagram of a curb data acquisition device provided in an embodiment of this application; Figure 6 This is a schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] The following disclosure provides numerous different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of the invention. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0027] To address the technical problem of the inability to obtain real-time and accurate roadside data in existing technologies, this application provides a roadside data acquisition method that can achieve accurate extraction of roadside data.

[0028] Figure 1 This is a flowchart illustrating a method for acquiring curb data according to an embodiment of this application. Figure 1 As shown, the above-mentioned method for obtaining curb data includes: S102, acquire target point cloud data of the moving tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the moving tool within the target frame; S104, divide the target point cloud data into a grid network of a preset area; S106, For each grid in the grid network, filter out the ground laser point cloud in the grid to obtain the remaining laser point cloud; S108, classify the grid including the remaining laser point cloud according to the gradient information to obtain the road edge data of the target frame; S110, merge the roadside data of the target frame with the roadside data of the historical frame to obtain the roadside data of the road where the mobile vehicle is located.

[0029] The roadside data acquisition method provided in this application can be applied to various scenarios of autonomous driving or assisted driving. By providing accurate roadside data for mobile vehicles, it enables mobile vehicles to work or move precisely. Mobile vehicles can be any equipment capable of mobility, including vehicles with autonomous or intelligent driving capabilities (including passenger vehicles (such as cars, buses, coaches, minibuses, etc.), cargo vehicles (such as ordinary trucks, box trucks, trailer trucks, enclosed trucks, tank trucks, flatbed trucks, container trucks, dump trucks, special structure trucks), special vehicles (such as logistics delivery vehicles, automated guided vehicles (AGVs), patrol vehicles, cranes, excavators, bulldozers, loaders, road rollers, off-road engineering vehicles, armored engineering vehicles, sewage treatment vehicles, sanitation vehicles, vacuum trucks, floor scrubbers, water sprinkler trucks, sweeping robots, food delivery robots, shopping guide robots, lawnmowers, golf carts, etc.), recreational vehicles (such as amusement vehicles, amusement park autonomous driving devices, balance bikes, etc.), rescue vehicles (such as fire trucks, ambulances, power repair vehicles, engineering emergency rescue vehicles, etc.)), and robots (such as sweeping robots, food delivery robots, etc.).

[0030] The curb data in this application may include the position and height data of the curb relative to the mobile tool, and may also include other data, which are not limited in this application. The purpose of acquiring curb data is to enable the mobile tool to use the curb data as a reference during automatic movement, thereby making its movement or working process accurate.

[0031] During the movement of a mobile vehicle, LiDAR can acquire laser point cloud data within a certain range in front of the vehicle. When acquiring laser point cloud data, the LiDAR obtains the data frame by frame; that is, within one second, the current laser point cloud data is acquired once per frame. Combining these multiple frames constitutes the total laser point cloud data for the entire movement of the mobile vehicle. In this example, after the LiDAR acquires the laser point cloud data, the laser point cloud data of a predetermined area in front of the mobile vehicle in each frame can be extracted as the target point cloud data. This allows for the calculation of the roadside location for the mobile vehicle at each frame, and by combining the roadside data from multiple frames, the roadside data can be comprehensively determined.

[0032] By performing the following processing on each frame as the target frame, the path edge data for each frame can be determined: The laser point cloud data of a preset area in front of the moving tool is acquired in the target frame and used as the target point cloud data. Then, the preset area is divided into a grid network composed of multiple grids (or meshes), and the grid network to which the target point cloud data belongs is determined. Thus, the target point cloud data is divided into different grids under the grid network. Each grid includes laser point clouds of multiple target point cloud data. Each laser point cloud has a corresponding position within the grid, which includes horizontal and vertical coordinates. The horizontal and vertical coordinates represent the position of the laser point cloud mapped to the ground, and the vertical coordinate represents the height of the laser point cloud.

[0033] For each grid cell containing laser point clouds, the ground laser point clouds are identified and then deleted to obtain the remaining point cloud data. This remaining point cloud data consists of all point cloud data except for the ground laser point clouds, with the majority being roadside laser point clouds. If obstacle interference is removed, the remaining laser point clouds are almost entirely composed of roadside laser point clouds. The grid cells containing these laser point clouds are then categorized according to gradient information to determine the grid cells containing the roadside, thus obtaining information such as the roadside's position and height, and ultimately, the roadside data for the target frame.

[0034] After obtaining the curb data for each frame, the curb data of each frame is combined and judged to determine the curb data of the road where the mobile vehicle is located. Whether the mobile vehicle is driving autonomously or cleaning automatically, it can perform its operations based on the curb data.

[0035] The solution provided in this application embodiment acquires target point cloud data of a mobile tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile tool within a target frame; divides the target point cloud data into a grid network of the preset area; filters out the ground laser point cloud in each grid cell to obtain the remaining laser point cloud; classifies the grid cells including the remaining laser point cloud according to gradient information to obtain roadside data of the target frame; and merges the roadside data of the target frame with the roadside data of historical frames to obtain the roadside data of the road where the mobile tool is located. Thus, accurate roadside data can be extracted based on the laser point cloud data in the preset area in front of the mobile tool, solving the technical problem in the prior art that it is impossible to obtain real-time and accurate roadside data.

[0036] To avoid interference from noisy point cloud data in the target point cloud data, before filtering out the ground laser point cloud in each grid of the grid network to obtain the remaining laser point cloud, obstacles on the road where the mobile vehicle is located can be determined; the target position of the obstacle within the target frame can be determined; and the grid corresponding to the target position in the grid network can be masked.

[0037] This embodiment describes the step of masking the obstacle laser point cloud after dividing the target point cloud data in the target frame into different grids of a grid network. Masking the obstacle laser point cloud can remove the interference of obstacles near the curb on the determination process of the curb data.

[0038] When creating a laser point cloud to mask obstacles, the first step is to identify the obstacles. There are several methods for identifying obstacles. For example, images can be captured, and moving objects in different frames can be identified; these moving objects are then considered obstacles. Alternatively, a recognition model can be used to identify objects outside the curb as obstacles.

[0039] After the obstacles are identified, the interference of the grid cells containing the obstacles needs to be removed from the target point cloud data of the current frame. For the remaining grid cells, the process of acquiring the road edge data is then performed.

[0040] For example, taking an automated cleaning vehicle as an example, a camera can be installed on the vehicle to capture images of the area in front of it. For different frames of the image, the camera identifies identical objects within the images and determines the object's position in real space for each frame. If a change in the object's position is detected, it is determined that the object is moving, and further identified as an obstacle. Alternatively, the automated cleaning vehicle's camera can capture images of the area in front of it, and after capturing the images, it can identify all objects in the images, classifying objects such as trash cans, bicycles, and trees—objects not located on the curb—as obstacles.

[0041] After identifying obstacles, the area containing the obstacle is enclosed by a special frame to indicate that the object is an obstacle. After acquiring the target point cloud data of the target frame and dividing the target point cloud data into different grids according to the grid network, the grids that belong to obstacles can be masked, and the curb data can be determined for the remaining grids.

[0042] This embodiment identifies obstacles near the curb and removes the corresponding grid cells in the grid network, thereby removing obstacle interference during the curb data acquisition process and extracting accurate curb data.

[0043] If the target point cloud data is to be divided into a grid network of a preset area, one possible method is to divide the preset area into multiple blocks according to the plane where the road surface is located. The size of each block varies depending on its distance from the moving vehicle. The block in which each target point cloud data is located is then determined, thus completing the division of the target point cloud data.

[0044] This embodiment provides a method for dividing target point cloud data of a target frame into a raster network. Specifically, multiple methods can be used for division. The purpose of division is to break down the entire preset area into smaller raster networks, thereby dividing a large area into smaller areas for operation, improving the precision of the operation. The division method can be equal division, that is, dividing the preset area into multiple completely equal raster cells, or uneven division, such as determining the raster size according to the distance from the moving tool; the closer to the moving tool, the smaller the raster, thus allowing for more precise data processing. One division method is to divide the preset area with a straight line based on the distance from the moving tool, with one side of the line representing the closer side and the other side representing the farther side. The closer side is divided into a finer raster network, and the farther side is divided into a slightly coarser raster network, allowing the moving tool to extract more detailed curb data in the closer area and slightly less detailed curb data in the farther area. Another method of partitioning the area involves using a point on the moving vehicle as the origin, and the direction of the vehicle's heading and the direction perpendicular to the heading as the horizontal and vertical coordinate axes. Along one direction of the horizontal and vertical coordinate axes, the predefined area is divided into equal sub-regions; along the other direction of the horizontal and vertical coordinate axes, each sub-region is further divided into multiple blocks. In this scheme, any point on the moving vehicle can be chosen as the origin, or a point on the axis of symmetry can be selected, such as the point on the axis of symmetry closest to the end of the moving vehicle. After determining the origin, using the two lines along the direction of the vehicle's heading and the direction perpendicular to the heading as the horizontal and vertical coordinates constructs a two-dimensional region of the ground plane in a three-dimensional coordinate system. Within this two-dimensional region, the farther the vehicle's heading is from the moving vehicle, the smaller the grid size.

[0045] This example demonstrates how dividing the grid network by varying the size of the blocks based on their distance from the moving vehicle allows for the creation of a more detailed grid network for areas closer to the moving vehicle, thereby improving the accuracy of acquiring roadside data near the moving vehicle.

[0046] If we want to filter out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud, we can perform the following operations on each grid cell: fit the laser point cloud with a height difference less than a preset value into a plane; filter out the laser point cloud between the plane and the preset distance above the plane to obtain the remaining laser point cloud.

[0047] This example provides a method for filtering out ground laser point clouds in a grid. The purpose is to delete the ground and near-ground laser point clouds from all laser point clouds in the grid. The remaining laser point clouds are mostly roadside laser point clouds. If the grid has already removed obstacles in the above steps, the remaining point clouds are almost all roadside laser point clouds.

[0048] To remove laser point clouds from the ground, the first step is to determine which point clouds constitute the ground. In this example, there are several ways to determine ground laser point clouds. One method is to fit laser point clouds within a grid with height differences less than a preset value to a plane, and then consider this plane as the ground. This method leverages the flatness of the ground to identify point clouds that match ground features from the grid's point cloud data, thus identifying the ground. Since curbs may also have complete planes, if the laser point clouds are fitted to multiple planes, the plane with the lowest height is identified as the ground. Another method for identifying the ground is to obtain the number of laser point clouds within the grid whose height is lower than a certain threshold, and then determine the proportion of these laser point clouds to the total number of laser point clouds within the grid. If the proportion is low, the threshold is increased until the number of obtained laser point clouds reaches a preset condition. At this point, the obtained laser point clouds are considered ground laser point clouds.

[0049] After identifying the laser point cloud on the ground, the laser point cloud on the ground and the laser point cloud at a preset distance above the ground are deleted, thereby removing the interference of the laser point cloud on the ground and near the ground, while retaining the laser point cloud on the roadside.

[0050] This example fits the laser point cloud within the grid with a height difference less than a preset value to a plane; filters out the laser point cloud between the plane and a preset distance above the plane to obtain the remaining laser point cloud. This allows for accurate determination of the ground laser point cloud within the grid, enabling its filtering and retention of accurate roadside laser point clouds.

[0051] If the grid including the remaining laser point cloud is classified according to the gradient information to obtain the path edge data of the target frame, the grid including the remaining laser point cloud can be traversed to calculate the vertical gradient of the grid; the grids with vertical gradients greater than the preset gradient and adjacent grids are merged into one connected grid; the merged one or more connected grids are determined as the path edge data of the target frame.

[0052] In this example, if the grid cells have already had obstacles removed, and each remaining grid cell has been processed to remove ground laser point clouds, then the remaining laser point clouds are almost entirely roadside laser point clouds. At this point, different grid sets can be merged. The purpose of merging is to make a final judgment and filter out the laser point clouds from the remaining laser point clouds that can form a complete roadside laser point cloud. The extraction method involves calculating the vertical gradient of the grid cell using the remaining laser point clouds within the grid after obstacle removal. The larger the vertical gradient, the more drastic the change in the laser point cloud in the vertical direction, indicating that this part of the laser point cloud is a roadside laser point cloud. Merging grid cells with vertical gradients greater than a preset gradient yields coherent roadside data.

[0053] It should be noted that during merging, only adjacent grid cells with a vertical gradient greater than the preset gradient are merged. If grid cells are neither directly nor indirectly adjacent, merging cannot be performed, resulting in multiple road edges, indicating that there are breaks or damage to the road edges. In this case, virtual laser point clouds can be added at the breaks or damage points according to the pattern of the multiple road edges to assist the movement or operation of the mobile tool.

[0054] This embodiment merges scattered roadside data by merging the grid according to the vertical gradient, thereby obtaining complete roadside data.

[0055] If the road edge data of the target frame is merged with the road edge data of the historical frame to obtain the road edge data of the road where the mobile tool is located, the following method can be used: when the number of times a certain area is determined to be a road edge in the road edge data of the target frame and the road edge data of the historical frame exceeds a preset number, a certain area is determined to be a road edge; after determining whether each area is a road edge, the road edge data of the road is obtained.

[0056] In this example, after obtaining the curb edge data for each frame in the previous example, the process of fusing curb edge data from multiple frames is described. When fusing curb edge data from multiple frames, it's necessary to determine whether the judgment results for the same region are the same in different frames. For example, if the same region is judged as a curb edge in different frames, then that region is determined to be a curb edge. If the same region is judged as a curb edge in some frames and not in others, a comprehensive judgment is needed. Higher weights can be assigned to adjacent frames and lower weights to historical frames to obtain a comprehensive result. If the comprehensive result indicates that the region is not a curb edge, then that region is determined not to be a curb edge. Figure 2 As shown, Figure 2 In the process, the roadside data, which are indicated as roadside edges in the comprehensive results of different frames, are merged. Figure 2 (Only a few frames are shown), which yields the curb data for the entire road.

[0057] This example obtains complete and accurate roadside data by comprehensively judging the roadside data of all frames, which can be used for the movement or operation of mobile tools.

[0058] This example illustrates how an automated cleaning vehicle moves along the roadside. As the vehicle moves slowly forward, the LiDAR system extracts multiple frames of laser point cloud data from in front of the vehicle every second, saving each frame. In this example, each frame of laser point cloud data is obtained from the LiDAR and then processed to obtain the roadside data for that frame. The specific processing steps are as follows: Figure 3 As shown.

[0059] 1. Extract the laser point cloud data of a predetermined area in front of the vehicle from a frame of laser point cloud data from the LiDAR, and use this as a frame of laser point cloud data. In other words, acquire the predetermined laser point cloud data in front of the vehicle in each frame as the vehicle moves forward.

[0060] 2. Divide the preset region in each frame into a grid network, and determine which grid each laser point cloud belongs to within the grid network. The specific division method can be as follows: Figure 4 As shown, Figure 4 In this system, the center of the vehicle's rear axle is used as the origin. The direction of the vehicle's front is one coordinate axis, and the lateral direction, perpendicular to the front direction, is another coordinate axis, forming a two-dimensional coordinate system parallel to the ground. A grid network is constructed within this two-dimensional coordinate system. The Region of Interest (ROI) is defined as follows: the area [0, 12] meters in front of the vehicle is called the near end, and the area [12, 16] meters is called the far end. The area ±3 meters to the left and right of the vehicle is a preset region. The near-end grid is 0.1 meters wide and 0.1 meters long, and the far-end grid is 0.1 meters wide and 0.3 meters long, forming the grid network. Laser point clouds are filled into the grid according to their positions. The height difference of the laser point clouds within each grid and the average of their coordinate values ​​in the two-dimensional coordinate system can be calculated.

[0061] The grid also needs to be filtered. The grid corresponding to the Box position of the marked obstacle (vehicle, pedestrian and other moving target) is not included in the roadside extraction calculation. Therefore, the laser point cloud and statistical information of the grid at the corresponding position are filtered by the Box position information.

[0062] 3. For the remaining grid cells, remove the ground point cloud based on the feature information within the grid cells.

[0063] Cells with height differences less than a threshold of 0.03 are selected from the grid and used to fit the plane Ax+By+Cz+D=0. The fitted plane is the ground. The fitted plane is fitted using the least squares method. All grids are traversed, and point clouds within 0.05 meters above the fitted plane are filtered out and discarded, and not included in subsequent calculations.

[0064] For the remaining point cloud, the Sobel operator Gy is used to calculate the gradient information of each grid cell in the lateral direction of the vehicle: Gy=\begin{bmatrix}-1&-2&-1 \\0&0&0 \\ 1&2&1\end{bmatrix} The gradient information of the obtained raster is:

[0065] 4. Using the connected component algorithm, adjacent grid cells with gradients greater than the gradient threshold are classified as a single instance.

[0066] In this example, the 8-neighborhood operator is used to calculate and obtain connected component instances. Adjacent instances are merged into one instance, thus obtaining the curb data for one frame.

[0067] Multi-frame fusion branch calculation steps: 1) Transfer the edge data from the previous frame to the current frame using pose information.

[0068] 2) After matching and merging the path edge of the current frame laser point cloud computing connected domain with the path edge of the previous frame, output the path edge data of the current frame.

[0069] The automatic sweeper moves and sweeps the road based on the current curb data.

[0070] Figure 5 This is a schematic diagram of a curb data acquisition device provided in an embodiment of this application. Figure 5 As shown, the aforementioned curb data acquisition device includes: The acquisition module 502 is used to acquire target point cloud data of the moving tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the moving tool within the target frame; The partitioning module 504 is used to partition the target point cloud data into a grid network of a preset area; Processing module 506 is used to filter out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud; The classification module 508 is used to classify the grid including the remaining laser point cloud according to the gradient information to obtain the road edge data of the target frame. The merging module 510 is used to merge the roadside data of the target frame with the roadside data of the historical frame to obtain the roadside data of the road where the mobile vehicle is located.

[0071] The roadside data acquisition method provided in this application can be applied to various scenarios of autonomous driving or assisted driving. By providing accurate roadside data for mobile vehicles, it enables mobile vehicles to work or move precisely. Mobile vehicles can be any equipment capable of mobility, including vehicles with autonomous or intelligent driving capabilities (including passenger vehicles (such as cars, buses, coaches, minibuses, etc.), cargo vehicles (such as ordinary trucks, box trucks, trailer trucks, enclosed trucks, tank trucks, flatbed trucks, container trucks, dump trucks, special structure trucks), special vehicles (such as logistics delivery vehicles, automated guided vehicles (AGVs), patrol vehicles, cranes, excavators, bulldozers, loaders, road rollers, off-road engineering vehicles, armored engineering vehicles, sewage treatment vehicles, sanitation vehicles, vacuum trucks, floor scrubbers, water sprinkler trucks, sweeping robots, food delivery robots, shopping guide robots, lawnmowers, golf carts, etc.), recreational vehicles (such as amusement vehicles, amusement park autonomous driving devices, balance bikes, etc.), rescue vehicles (such as fire trucks, ambulances, power repair vehicles, engineering emergency rescue vehicles, etc.)), and robots (such as sweeping robots, food delivery robots, etc.).

[0072] The curb data in this application may include the position and height data of the curb relative to the mobile tool, and may also include other data, which are not limited in this application. The purpose of acquiring curb data is to enable the mobile tool to use the curb data as a reference during automatic movement, thereby making its movement or working process accurate.

[0073] During the movement of a mobile vehicle, LiDAR can acquire laser point cloud data within a certain range in front of the vehicle. When acquiring laser point cloud data, the LiDAR obtains the data frame by frame; that is, within one second, the current laser point cloud data is acquired once per frame. Combining these multiple frames constitutes the total laser point cloud data for the entire movement of the mobile vehicle. In this example, after the LiDAR acquires the laser point cloud data, the laser point cloud data of a predetermined area in front of the mobile vehicle in each frame can be extracted as the target point cloud data. This allows for the calculation of the roadside location for the mobile vehicle at each frame, and by combining the roadside data from multiple frames, the roadside data can be comprehensively determined.

[0074] By performing the following processing on each frame as the target frame, the path edge data for each frame can be determined: The laser point cloud data of a preset area in front of the moving tool is acquired in the target frame and used as the target point cloud data. Then, the preset area is divided into a grid network composed of multiple grids (or meshes), and the grid network to which the target point cloud data belongs is determined. Thus, the target point cloud data is divided into different grids under the grid network. Each grid includes laser point clouds of multiple target point cloud data. Each laser point cloud has a corresponding position within the grid, which includes horizontal and vertical coordinates. The horizontal and vertical coordinates represent the position of the laser point cloud mapped to the ground, and the vertical coordinate represents the height of the laser point cloud.

[0075] For each grid cell containing laser point clouds, the ground laser point clouds are identified and then deleted to obtain the remaining point cloud data. This remaining point cloud data consists of all point cloud data except for the ground laser point clouds, with the majority being roadside laser point clouds. If obstacle interference is removed, the remaining laser point clouds are almost entirely composed of roadside laser point clouds. The grid cells containing these laser point clouds are then categorized according to gradient information to determine the grid cells containing the roadside, thus obtaining information such as the roadside's position and height, and ultimately, the roadside data for the target frame.

[0076] After obtaining the curb data for each frame, the curb data of each frame is combined and judged to determine the curb data of the road where the mobile vehicle is located. Whether the mobile vehicle is driving autonomously or cleaning automatically, it can perform its operations based on the curb data.

[0077] The solution provided in this application embodiment acquires target point cloud data of a mobile tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile tool within a target frame; divides the target point cloud data into a grid network of the preset area; filters out the ground laser point cloud in each grid cell to obtain the remaining laser point cloud; classifies the grid cells including the remaining laser point cloud according to gradient information to obtain roadside data of the target frame; and merges the roadside data of the target frame with the roadside data of historical frames to obtain the roadside data of the road where the mobile tool is located. Thus, accurate roadside data can be extracted based on the laser point cloud data in the preset area in front of the mobile tool, solving the technical problem in the prior art that it is impossible to obtain real-time and accurate roadside data.

[0078] Optionally, the above device further includes: a shielding module, used to determine obstacles on the road where the mobile tool is located before filtering out the ground laser point cloud in each grid of the grid network to obtain the remaining laser point cloud; determine the target position of the obstacle within the target frame; and shield the grid corresponding to the target position in the grid network.

[0079] Optionally, the above-mentioned segmentation module includes: a segmentation unit, used to divide the preset area into multiple blocks according to the plane where the road surface is located, wherein the size of the blocks varies depending on the distance from the moving tool; and to determine the block where each target point cloud data is located, thereby completing the segmentation of the target point cloud data.

[0080] Optionally, the above processing module includes: a processing unit, used to perform the following operations on each grid in the grid network: fitting the laser point cloud with a height difference less than a preset value within the grid to a plane; filtering out the laser point cloud between the plane and a preset distance above the plane to obtain the remaining laser point cloud.

[0081] Optionally, the above classification module includes: a classification unit, used to traverse the grid including the remaining laser point cloud, calculate the vertical gradient of the grid; merge adjacent grids with a vertical gradient greater than a preset gradient into one connected grid; and determine the merged one or more connected grids as the path edge data of the target frame.

[0082] Optionally, the above merging module includes: a merging unit, used to determine a certain area as a road edge when the number of times a certain block is determined to be a road edge in the road edge data of the target frame and the road edge data of the historical frame exceeds a preset number; and to obtain the road edge data of the road after determining whether each area is a road edge.

[0083] For other examples of this embodiment, please refer to the examples above, which will not be repeated here.

[0084] like Figure 6 As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. Memory 113 is used to store computer programs; In one embodiment of this application, the processor 111, when executing the program stored in the memory 113, implements the curb data acquisition method provided in any of the foregoing method embodiments.

[0085] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the curb data acquisition method provided in any of the foregoing method embodiments.

[0086] This application also provides a mobile tool, including the electronic device described above.

[0087] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0088] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0089] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0090] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for acquiring curb data, characterized in that, include: Acquire target point cloud data of a mobile tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile tool within a target frame; The target point cloud data is divided into a grid network of the preset region; For each grid cell in the grid network, the ground laser point cloud in the grid cell is filtered out to obtain the remaining laser point cloud; The grid including the remaining laser point cloud is classified according to gradient information to obtain the path edge data of the target frame; The roadside data of the target frame is merged with the roadside data of the historical frames to obtain the roadside data of the road where the mobile vehicle is located.

2. The method according to claim 1, characterized in that, Before filtering out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud, the method further includes: Identify obstacles on the road where the mobile vehicle is located; Determine the target position of the obstacle within the target frame; The grid corresponding to the target location in the grid network is masked.

3. The method according to claim 1, characterized in that, The step of dividing the target point cloud data into the grid network of the preset region includes: The preset area is divided into multiple blocks according to the plane of the road surface, wherein the size of the blocks varies depending on their distance from the mobile tool; The block containing each target point cloud data is determined, thus completing the division of the target point cloud data.

4. The method according to claim 1, characterized in that, The step of filtering out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud includes: For each cell in the grid network, perform the following operation: Within the grid, the laser point cloud with a height difference less than a preset value is fitted into a plane; The laser point cloud between the plane and a preset distance above the plane is filtered out to obtain the remaining laser point cloud.

5. The method according to claim 1, characterized in that, The process of classifying the grid including the remaining laser point cloud according to gradient information to obtain the path edge data of the target frame includes: The grid including the remaining laser point cloud is traversed, and the vertical gradient of the grid is calculated. Merge adjacent grid cells with a vertical gradient greater than the preset gradient into one connected grid cell; One or more connected graticles that are merged are identified as the curb data of the target frame.

6. The method according to claim 1, characterized in that, The step of merging the roadside data of the target frame with the roadside data of historical frames to obtain the roadside data of the road where the mobile vehicle is located includes: When the number of times a certain region is identified as a road edge in the road edge data of the target frame and the road edge data of the historical frame exceeds a preset number, the certain region is identified as a road edge. Once it is determined whether each area is a curb, the curb data of the road is obtained.

7. A device for acquiring curb data, characterized in that, include: The acquisition module is used to acquire target point cloud data of the mobile tool, wherein the target point cloud data is laser point cloud data of a preset area in front of the mobile tool within the target frame; A partitioning module is used to partition the target point cloud data into a grid network of the preset region; The processing module is used to filter out the ground laser point cloud in each grid cell of the grid network to obtain the remaining laser point cloud; The classification module is used to classify the grid including the remaining laser point cloud according to gradient information to obtain the path edge data of the target frame; The merging module is used to merge the roadside data of the target frame with the roadside data of the historical frame to obtain the roadside data of the road where the mobile vehicle is located.

8. An electronic device, characterized in that, include: At least one communication interface; At least one bus connected to the at least one communication interface; At least one processor connected to the at least one bus; At least one memory connected to the at least one bus, wherein the memory stores a computer program, and the processor executes the computer program to implement the curb data acquisition method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The storage medium stores computer-executable instructions for performing the curb data acquisition method according to any one of claims 1 to 6 of this application.

10. A mobile tool, characterized in that, Includes the electronic device shown in claim 9.

Images