Point cloud data motion segmentation method, device, computer equipment and storage medium

Abstract

A point cloud data motion segmentation method comprises: acquiring multiple frames of point cloud data; performing point cloud motion segmentation on the point cloud data to obtain point cloud motion segmentation results; performing ground filtering processing on each frame of point cloud data to obtain multiple frames of non-ground point cloud data; performing cluster analysis on the multiple frames of non-ground point cloud data to obtain three-dimensional bounding boxes corresponding to each frame of non-ground point cloud data; performing occlusion analysis on the three-dimensional bounding box to determine the point cloud motion state corresponding to the three-dimensional bounding box; when the point cloud motion state is dynamic, updating the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain a target point cloud motion segmentation result.

Description

Technical Field

[0001] The present application relates to a point cloud data motion segmentation method, device, computer equipment and storage medium. Background Art

[0002] LiDAR sensors can provide real-time and accurate three-dimensional scene information, with a large ranging range, high accuracy, and the measurement value is not affected by ambient light. They are widely used in unmanned driving, security monitoring, surveying and exploration, and other fields. The information collected by LiDAR sensors is usually presented in the form of point clouds. During the unmanned driving process, in order to ensure the safety of unmanned vehicles, it is necessary to monitor objects in the surrounding environment in real time, especially to detect and track moving objects such as pedestrians and moving vehicles. Therefore, it is necessary to segment the point cloud data to identify moving objects.

[0003] In the traditional point cloud motion segmentation method, non-deep learning methods are used, such as based on grid maps, to count the change status of point cloud data in the grid corresponding to multiple consecutive frames, so as to identify dynamic points and static points in the point cloud data. However, the traditional method only considers the individual features of each point, and it is difficult to directly correspond the point cloud data of consecutive frames, resulting in low accuracy of point cloud motion segmentation. Summary of the invention

[0004] According to various embodiments disclosed in the present application, a point cloud data segmentation method, apparatus, computer device, and storage medium are provided that can improve the accuracy of point cloud motion segmentation.

[0005] A point cloud data motion segmentation method, comprising:

[0006] Get multi-frame point cloud data;

[0007] Perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0008] Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0009] Perform cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0010] Performing occlusion analysis on the three-dimensional bounding box to determine the motion state of the point cloud corresponding to the three-dimensional bounding box; and

[0011] When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0012] A point cloud data motion segmentation device, comprising:

[0013] Communication module, used to obtain multi-frame point cloud data;

[0014] The point cloud motion segmentation module is used to perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0015] A ground filtering module is used to perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0016] A cluster analysis module is used to perform cluster analysis on multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0017] An occlusion analysis module, used to perform occlusion analysis on the three-dimensional bounding box to determine the motion state of the point cloud corresponding to the three-dimensional bounding box; and

[0018] The updating module is used to update the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box when the point cloud motion state is dynamic, so as to obtain the target point cloud motion segmentation result.

[0019] A computer device includes a memory and one or more processors, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the one or more processors perform the following steps:

[0020] Get multi-frame point cloud data;

[0021] Perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0022] Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0023] Perform cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0024] Performing occlusion analysis on the three-dimensional bounding box to determine the motion state of the point cloud corresponding to the three-dimensional bounding box; and

[0025] When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0026] One or more non-volatile computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps:

[0027] Get multi-frame point cloud data;

[0028] Perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0029] Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0030] Perform cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0031] Performing occlusion analysis on the three-dimensional bounding box to determine the motion state of the point cloud corresponding to the three-dimensional bounding box; and

[0032] When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0033] The details of one or more embodiments of the present application are set forth in the following drawings and description. Other features and advantages of the present application will become apparent from the description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A diagram of an application environment of a point cloud data motion segmentation method in one or more embodiments.

[0036] Figure 2 A schematic diagram of a flow chart of a point cloud data motion segmentation method in one or more embodiments.

[0037] Figure 3 A flowchart of the steps of performing occlusion analysis on a three-dimensional bounding box and determining the motion state of a point cloud corresponding to the three-dimensional bounding box in one or more embodiments.

[0038] Figure 4 A block diagram of a point cloud data motion segmentation device in one or more embodiments.

[0039] Figure 5 is a block diagram of a computer device in accordance with one or more embodiments. DETAILED DESCRIPTION

[0040] In order to make the technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] The point cloud data motion segmentation method provided in this application can be applied to Figure 1 In the application environment shown. The vehicle-mounted sensor 102 sends the collected multi-frame point cloud data to the computer device 104. For example, the vehicle-mounted sensor can be a laser radar. The computer device 104 performs point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result. The computer device 104 can also perform ground filtering processing on each frame of point cloud data to obtain multiple frames of non-ground point cloud data, perform clustering analysis on the multiple frames of non-ground point cloud data, and obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data, thereby performing occlusion analysis on the three-dimensional bounding box to determine the point cloud motion state corresponding to the three-dimensional bounding box. When the point cloud motion state is dynamic, the computer device 104 updates the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain a target point cloud motion segmentation result.

[0042] In one embodiment, if Figure 2 As shown, a point cloud data motion segmentation method is provided, which is applied to Figure 1 The computer device in the example is used to illustrate, including the following steps:

[0043] Step 202, obtaining multiple frames of point cloud data.

[0044] Point cloud data can be data recorded in the form of point clouds by the on-board sensor scanning the surrounding environment information. For example, the on-board sensor can be a laser radar. Single-frame point cloud data refers to the original point cloud collected by the on-board sensor rotating horizontally for one circle. Multi-frame point cloud data refers to the original point cloud of multiple consecutive frames in a temporal order. The original point cloud is distributed on multiple circular scanning lines with different vertical angles, and the point cloud data on each circular scanning line is composed of a circle of laser points. Point cloud data can specifically include the three-dimensional coordinates (x, y, z) of each point, laser reflection intensity (Intensity), color information (RGB), etc. The three-dimensional coordinates are used to represent the position information of the surface of the target object in the surrounding environment. For example, the three-dimensional coordinates can be the coordinates of a point in a Cartesian coordinate system, specifically including the horizontal axis coordinates, vertical axis coordinates and vertical axis coordinates of the point in the Cartesian coordinate system. The Cartesian coordinate system is a three-dimensional space coordinate system established with the laser radar as the origin, and the three-dimensional space coordinate system includes the horizontal axis (x axis), the vertical axis (y axis) and the vertical axis (z axis). The three-dimensional space coordinate system established with the laser radar as the origin satisfies the right-hand rule. The x-axis coordinate in the three-dimensional coordinate represents the longitudinal distance of the scanned target object surface relative to the laser radar, the y-axis coordinate represents the lateral offset of the scanned target object surface relative to the laser radar, and the z-axis coordinate represents the height of the scanned target object surface relative to the laser radar.

[0045] During the unmanned driving process, the vehicle can scan the surrounding environment through the laser radar installed on the vehicle to obtain corresponding multi-frame point cloud data, and transmit the collected multi-frame point cloud data to the computer equipment.

[0046] Step 204 , performing point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result.

[0047] Point cloud motion segmentation refers to the motion classification of points in point cloud data to identify whether each point is a dynamic point or a static point. After obtaining multiple frames of point cloud data transmitted by the vehicle-mounted sensor, the computer device can perform point cloud motion segmentation on the point cloud data. The point cloud motion segmentation result includes the motion state corresponding to each point in the point cloud data, including dynamic points and static points.

[0048] The method of performing point cloud motion segmentation on point cloud data can be to first divide the data space corresponding to each frame of point cloud data into multiple grids according to a preset grid size to obtain a grid map, so that the computer device counts the quantity change values ​​of point cloud data under each grid of multiple consecutive frames, and determines the points corresponding to the grids whose quantity change values ​​are greater than the threshold as dynamic points, and determines the points corresponding to the grids whose quantity change values ​​are less than or equal to the threshold as static points. In one embodiment, the computer device can also establish a spatial grid containing point cloud data, set each grid to a preset size, such as 0.1 meters, so as to divide the spatial grid into multiple grids. When the position corresponding to a grid has one or more points in the point cloud data, the grid can be marked as occupied. When there is no point at the position corresponding to the grid, the grid is marked as unoccupied, thereby marking each grid in the spatial grid. By marking the grids, the grids where the target object exists can be quickly determined, so as to perform point cloud motion segmentation. In one embodiment, the smaller the size of the grid, the higher the accuracy and resolution. That is, for the same spatial grid, the more grids there are, the higher the resolution of the spatial grid. The higher the resolution of the spatial grid, the higher the accuracy of the grid map.

[0049] In one of the embodiments, the point cloud motion segmentation method can be any one of a plurality of point cloud motion segmentation methods such as a point-level occlusion analysis method, a point cloud motion segmentation method based on a deep learning model, etc.

[0050] In one of the embodiments, point cloud motion segmentation is performed on point cloud data to obtain the point cloud motion segmentation result, including: calling a pre-trained point cloud motion segmentation model and inputting the point cloud data into the point cloud motion segmentation model; classifying the point cloud data through the point cloud motion segmentation model, outputting the motion state corresponding to each point in the point cloud data, and generating the point cloud motion segmentation result.

[0051] The point cloud motion segmentation model is obtained by training with a large amount of labeled sample data. For example, the point cloud motion segmentation model can be any one of pointnet, pointnet++, etc. The computer device can input the point cloud data into the point cloud motion segmentation model, classify the point cloud data by motion, and determine the motion state of each point in the point cloud data, that is, output whether each point is a dynamic point or a static point. The point cloud motion segmentation accuracy can be improved by classifying the point cloud data by motion using the pre-established point cloud motion segmentation model.

[0052] In one embodiment, the computer device may also pre-process the point cloud data before the point cloud motion segmentation. The pre-processing method may include downsampling and filtering processing operations. Specifically, the computer device may downsample the point cloud data by voxel grid filtering, which can remove noise points, reduce the amount of redundant data, and improve the point cloud processing speed. During filtering, the details can be maintained in the depth direction, and the loss of details in the depth direction can be reduced. Thus, the computer device performs through-filtering processing on the downsampled point cloud data. Specifically, the depth direction data range of the point cloud data can be set, for example, between 0.6-1.0m, and the point clouds outside this range will be removed as interference points. Then, the computer device uses a statistical filtering method for the point cloud data after through-filtering, performs a statistical analysis on the neighborhood of each point, and removes the point clouds outside the standard range, thereby reducing isolated redundant point clouds. Specifically, the computer device may use a sparse outlier removal method to calculate the average distance from each point in the point cloud data to all neighboring points, determine the points whose average distance is outside the standard range as outliers, and remove them from the point cloud data. The standard range is defined by the global distance mean and variance. By removing outliers, the accuracy of point cloud motion segmentation can be improved.

[0053] Step 206 , performing ground filtering processing on each frame of point cloud data to obtain multiple frames of non-ground point cloud data.

[0054] Step 208 , cluster analysis is performed on the multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data.

[0055] Since the traditional point cloud motion segmentation algorithm only considers the individual features of each point and lacks the semantic information of the point, it is difficult to directly correspond the point cloud data of continuous frames. Therefore, the computer equipment can also perform three-dimensional target-level occlusion analysis on the point cloud data while performing point cloud motion segmentation on the point cloud data to assist the traditional point cloud motion segmentation method and improve the accuracy of point cloud motion segmentation.

[0056] Before the computer device performs occlusion analysis at the three-dimensional target level on the point cloud data, it is necessary to first perform ground filtering processing, cluster analysis, and identification of three-dimensional bounding boxes on each frame of point cloud data. Ground filtering processing refers to filtering out ground points in the point cloud data, and the remaining points are non-ground points. The computer device can identify the ground points in each frame of point cloud data by performing ground segmentation on each frame of point cloud data, thereby filtering the ground points in each frame of point cloud data to obtain each frame of non-ground point cloud data. Specifically, the computer device can first divide the point cloud area where the point cloud data is located into multiple sub-areas. The point cloud area refers to the three-dimensional data space where the point cloud data is located. The division method can be to grid the point cloud area, that is, to divide the horizontal plane formed by the point cloud area in the x-axis direction and the y-axis direction. The grid division method can be equal division or random division. For example, for a vehicle-mounted sensor with a visible range of 100m, the size of the horizontal plane in the scanning area of ​​the vehicle-mounted sensor is 100m*100m, then the point cloud area can be equally divided into 10*10 horizontal grids. For each sub-area obtained by division, the computer device can estimate the corresponding ground using the least squares method according to the preset plane equation, thereby obtaining the ground corresponding to each sub-area. For example, the preset plane equation can be a three-variable linear equation. The ground corresponding to each sub-area is embodied in the form of a three-variable linear equation. The computer device traverses the input point coordinates in the corresponding sub-area in the equation corresponding to the ground, calculates the distance between each point and the corresponding ground, and when the distance is less than the threshold, the point is determined as a ground point. When the distance is greater than or equal to the threshold, the point is determined as a non-ground point. The threshold refers to the distance threshold used to determine whether the point is a ground point. The computer device then filters the ground points to obtain non-ground points in each frame of point cloud data, thereby obtaining each frame of non-ground point cloud data.

[0057] In one of the embodiments, the ground segmentation method can be any one of a method based on depth image and geometric relationship, a normal vector method, an absolute height method, an average height method, and the like.

[0058] The computer device performs cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each target object in each frame of non-ground point cloud data. Cluster analysis refers to clustering the points corresponding to the same target object together and identifying the three-dimensional bounding box corresponding to each target object. Cluster analysis can include two steps: clustering and identifying the three-dimensional bounding box. For example, the clustering method can be any one of the clustering algorithms such as the connected domain clustering method and the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. The three-dimensional bounding box corresponding to the target object can be any one of the L-shape fitting method, the principal component analysis method, etc. The three-dimensional bounding box may include the center point coordinates, size, orientation, etc. of the target object. By identifying the three-dimensional bounding box corresponding to the target object, different target objects can be accurately distinguished, which is conducive to improving the accuracy of point cloud motion segmentation.

[0059] Step 210: Perform occlusion analysis on the three-dimensional bounding box to determine the point cloud motion state corresponding to the three-dimensional bounding box.

[0060] Occlusion analysis refers to identifying the motion state of point cloud data based on the characteristics of the laser beam. The occlusion characteristics of the laser beam mean that when multiple objects are on the same laser scanning band, the laser radar will directly measure the target object closest to the vehicle, but will not measure other target objects on the same laser scanning band, and there are no other target objects between the laser radar and the measured target object. Among them, the laser scanning band can be the scanning range of the laser beam emitted by the laser radar at a certain moment. Therefore, the computer device can use the characteristics of the laser beam emitted by the laser radar to perform occlusion analysis on the three-dimensional bounding box of the target object corresponding to multiple frames of non-ground point cloud data. Specifically, the computer device can perform occlusion analysis on each three-dimensional bounding box of the current frame and the middle position of the line connecting the three-dimensional bounding box of the historical frame to the vehicle. When the computer device identifies a three-dimensional bounding box as dynamic during the occlusion analysis, all points in the three-dimensional bounding box are determined as dynamic points.

[0061] Step 212, when the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0062] Since the dynamic points obtained through occlusion analysis are highly accurate, the computer device can update the point cloud motion segmentation results based on the dynamic points obtained through occlusion analysis, and can correct the points in the point cloud motion segmentation results where the dynamic points are incorrectly recognized to obtain the target point cloud segmentation results.

[0063] In this embodiment, the computer device performs point cloud motion segmentation on the acquired multi-frame point cloud data to obtain the point cloud motion segmentation result. At the same time, the computer device performs ground filtering processing on each frame of point cloud data to obtain valid point cloud data. Cluster analysis is performed on multi-frame non-ground point cloud data to cluster points belonging to the same target object and obtain three-dimensional bounding boxes corresponding to each target object. The computer device performs occlusion analysis on the three-dimensional bounding box and determines the points in the dynamic three-dimensional bounding box as dynamic points. Since the three-dimensional bounding box is a real three-dimensional target, the occlusion analysis of the three-dimensional bounding box avoids the situation where the points corresponding to the same target object have different motion states, and can obtain a more accurate point cloud motion state, thereby obtaining a more accurate dynamic point. Then the computer device updates the point cloud motion segmentation result according to the dynamic point, thereby improving the accuracy of the point cloud motion segmentation.

[0064] In one embodiment, if Figure 3 As shown, the steps of performing occlusion analysis on the three-dimensional bounding box and determining the motion state of the point cloud corresponding to the three-dimensional bounding box include:

[0065] Step 302: Determine in the three-dimensional bounding box a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame.

[0066] Step 304 , performing occlusion analysis on the current three-dimensional bounding box and the historical three-dimensional bounding box to identify the point cloud motion state corresponding to the current three-dimensional bounding box.

[0067] The computer device can determine the current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and the historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame in the three-dimensional bounding box obtained by cluster analysis. The historical frame can be a certain historical frame or multiple historical frames within a preset time period. The current three-dimensional bounding box and the historical three-dimensional bounding box can be one or more. Thus, the computer device performs occlusion analysis on the current three-dimensional bounding box and the historical three-dimensional bounding box. In one embodiment, the current three-dimensional bounding box and the historical three-dimensional bounding box are subjected to occlusion analysis, and the point cloud motion state corresponding to the current three-dimensional bounding box is identified, including: obtaining the point cloud scan line corresponding to each historical three-dimensional bounding box; identifying the positional relationship between the current three-dimensional bounding box and the point cloud scan line; and identifying the point cloud motion state corresponding to the current bounding box according to the positional relationship. Specifically, the computer device obtains the point cloud scan line corresponding to each historical three-dimensional bounding box in the historical frame. The point cloud scan line refers to the line formed by the historical three-dimensional bounding box and the vehicle obtained on the laser scanning band emitted by the laser radar. The point cloud scan line includes the scanning angle of the laser radar. The computer device thus identifies the positional relationship between the current 3D bounding box and each point cloud scan line. Since there is no other target object in the middle of the line formed by the historical 3D bounding box and the vehicle, when the computer device identifies that there is a 3D bounding box located in the middle of the point cloud scan line in the current 3D bounding box, the 3D bounding box can be determined as a dynamic target, and all points in the 3D bounding box are dynamic points. The computer device only needs to determine the 3D bounding box as a dynamic target and the points in the 3D bounding box as dynamic points when it detects that there is a 3D bounding box located in the middle of each point cloud scan line in the current 3D bounding box. It can quickly determine the motion state of the point cloud.

[0068] For example, at time tk, the laser radar scanned the target object z(n, 1), indicating that there is no other target object in the middle of the line connecting the vehicle to the target object z(n, 1), and at time t, the laser radar measured the target object z(m, 1) in the middle of the point cloud scan line formed by the vehicle to the target object z(n, 1) at time tk, indicating that the target object z(m, 1) is dynamic relative to time tk. Therefore, the computer device can determine the point in the three-dimensional bounding box corresponding to the target object z(m, 1) as a dynamic point.

[0069] In this embodiment, the computer device performs occlusion analysis on the current 3D bounding box and the historical 3D bounding box, and by utilizing the characteristics of the laser beam, can quickly and accurately determine the point cloud motion state. In addition, since the 3D bounding box is a real 3D target, performing occlusion analysis on the 3D bounding box avoids the situation where points corresponding to the same target object have different motion states, and can obtain more accurate dynamic points.

[0070] In one of the embodiments, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box, and the point cloud data segmentation result obtained includes: extracting the dynamic point cloud data corresponding to the three-dimensional bounding box; determining the point cloud motion segmentation result corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation result; when the determined point cloud motion segmentation result is a static point, replacing the determined point cloud motion segmentation result with a dynamic point.

[0071] Since the accuracy of the dynamic points obtained through occlusion analysis is high, the computer device can update the point cloud motion segmentation result according to the dynamic point cloud data corresponding to the dynamic three-dimensional bounding box. Specifically, the computer device extracts the dynamic point cloud data corresponding to the dynamic three-dimensional bounding box, and the dynamic point cloud data includes multiple dynamic points. The computer device thus determines the point cloud motion segmentation result corresponding to each dynamic point in the point cloud motion segmentation result. When the determined point cloud motion segmentation result is a static point, the computer device updates the point cloud motion segmentation result to a dynamic point, thereby improving the recognition accuracy of the dynamic points in the point cloud motion segmentation result, and then improving the detection accuracy of dynamic targets.

[0072] In one of the embodiments, ground filtering processing is performed on each frame of point cloud data to obtain multiple frames of non-ground point cloud data, including: dividing the point cloud area corresponding to each frame of point cloud data into multiple grids; calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculating the distance value between each point in the point cloud data of each grid and the corresponding ground; filtering points with distance values ​​less than a first threshold value to obtain multiple frames of non-ground point cloud data.

[0073] The point cloud area refers to the data space where each frame of point cloud data is located. The computer device can divide the point cloud area corresponding to each frame of point cloud data into grids to obtain multiple grids. Grid division refers to dividing the point cloud area in the x-axis direction and the y-axis direction, that is, dividing the point cloud area in the xy horizontal plane.

[0074] Specifically, the computer device can divide the data area corresponding to the extracted point cloud data in the x-axis direction and the y-axis direction according to the preset parameters. The preset parameters can be parameters for grid division of the data area where the extracted point cloud data is located. For example, the preset parameters can be length*width, indicating the length and width of each grid obtained after grid division. The length and width can be the same or different. The preset parameters can also be equal division. The heights of multiple grids are the same. When the point cloud area corresponding to each frame of point cloud data is grid-divided, the order of division in the x-axis direction and the y-axis direction is not limited. For example, the computer device can first divide the point cloud area corresponding to the extracted point cloud data in the x-axis direction according to the preset parameters, and then divide the point cloud area corresponding to the extracted point cloud data in the y-axis direction according to the preset parameters. It is also possible to first divide the point cloud area corresponding to the extracted point cloud data in the y-axis direction according to the preset parameters, and then divide the point cloud area corresponding to the extracted point cloud data in the x-axis direction according to the preset parameters.

[0075] The computer device obtains a preset plane equation, and calculates the ground corresponding to the point cloud data in each grid according to the preset plane equation and the least square method, thereby obtaining the ground corresponding to each grid. The ground corresponding to each grid is embodied in the form of a three-variable first-order equation. The computer device traverses the point coordinates of each point in the corresponding grid and inputs them into the equation corresponding to the ground, and calculates the distance value between each point and the corresponding ground. A first threshold for determining the category of the point is pre-stored in the computer device. The category of the point may include ground points and non-ground points. The computer device compares the distance value with the first threshold, and when the distance value is less than the first threshold, it indicates that the point is a ground point. When the distance value is greater than or equal to the first threshold, it indicates that the point is a non-ground point. The computer device filters the points whose distance values ​​are less than the first threshold, and the remaining points constitute non-ground point cloud data.

[0076] In this embodiment, the point cloud area corresponding to each frame of point cloud data is divided into multiple grids, so as to calculate the ground corresponding to the point cloud data in each grid, and the ground points are filtered by calculating the distance between each point and the corresponding ground. Since the grid division only needs to divide the data area corresponding to the point cloud data in the x-axis direction and the y-axis direction, it can quickly perform ground filtering processing when the computing resources of the computer equipment are limited and the real-time requirements are high in the unmanned driving mode.

[0077] In one of the embodiments, calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation includes: selecting the point with the smallest height value in the point cloud data corresponding to each grid; calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; extracting points whose height difference is less than a second threshold, and performing plane fitting on the selected points according to the preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

[0078] After the computer device divides the point cloud area corresponding to each frame of point cloud data into grids and obtains multiple grids, it can select the point with the smallest height value in the point cloud data corresponding to each grid, and calculate the height difference between each point in the point cloud data of the corresponding grid and the point with the smallest height value. The computer device pre-stores a second threshold for determining whether it is a plane fitting point. When the height difference is less than the second threshold, it indicates that the point corresponding to the height difference is a plane fitting point. The computer device compares the height difference with the second threshold, selects the point whose height difference is less than the second threshold, and performs plane fitting on the selected point according to the preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

[0079] In this embodiment, the computer device selects the plane fitting point by calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value, and comparing the height difference with the second threshold value. Since the probability that the point with the smallest height value is the ground point is the largest, by calculating the height difference between each point and the point with the smallest height value, the plane fitting point can be determined more accurately, thereby improving the ground estimation efficiency and thereby improving the ground filtering efficiency.

[0080] In one embodiment, if Figure 4 As shown, a point cloud data motion segmentation device is provided, including: a communication module 402, a point cloud motion segmentation module 404, a ground filtering module 406, a clustering analysis module 408, an occlusion analysis module 410 and an updating module 412, wherein:

[0081] The communication module 402 is used to obtain multiple frames of point cloud data.

[0082] The point cloud motion segmentation module 404 is used to perform point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result.

[0083] The ground filtering module 406 is used to perform ground filtering processing on each frame of point cloud data to obtain multiple frames of non-ground point cloud data.

[0084] The cluster analysis module 408 is used to perform cluster analysis on multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data.

[0085] The occlusion analysis module 410 is used to perform occlusion analysis on the three-dimensional bounding box to determine the point cloud motion state corresponding to the three-dimensional bounding box.

[0086] The updating module 412 is used to update the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box when the point cloud motion state is dynamic, so as to obtain the target point cloud motion segmentation result.

[0087] In one of the embodiments, the occlusion analysis module 410 is also used to determine the current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and the historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame in the three-dimensional bounding box; perform occlusion analysis on the current three-dimensional bounding box and the historical three-dimensional bounding box to identify the point cloud motion state corresponding to the current three-dimensional bounding box.

[0088] In one of the embodiments, the occlusion analysis module 410 is further used to obtain the point cloud scan lines corresponding to each historical three-dimensional bounding box; identify the positional relationship between the current three-dimensional bounding box and the point cloud scan lines; and identify the point cloud motion state corresponding to the current bounding box based on the positional relationship.

[0089] In one of the embodiments, the update module 412 is also used to extract dynamic point cloud data corresponding to the three-dimensional bounding box; determine the point cloud motion segmentation results corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation results; when the determined point cloud motion segmentation results are static points, replace the determined point cloud motion segmentation results with dynamic points.

[0090] In one of the embodiments, the ground filtering module 406 is also used to divide the point cloud area corresponding to each frame of point cloud data into multiple grids; calculate the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculate the distance value between each point in the point cloud data of each grid and the corresponding ground; filter the points whose distance value is less than a first threshold value to obtain multiple frames of non-ground point cloud data.

[0091] In one of the embodiments, the ground filtering module 406 is also used to select the point with the smallest height value in the point cloud data corresponding to each grid; calculate the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; extract the points whose height difference is less than the second threshold value, and perform plane fitting on the selected points according to the preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

[0092] In one of the embodiments, the point cloud motion segmentation module 404 is also used to call a pre-trained point cloud motion segmentation model and input the point cloud data into the point cloud motion segmentation model; classify the point cloud data through the point cloud motion segmentation model, output the motion state corresponding to each point in the point cloud data, and generate a point cloud motion segmentation result.

[0093] For the specific definition of the point cloud data motion segmentation device, please refer to the definition of the trajectory prediction method above, which will not be repeated here. Each module in the above-mentioned point cloud data motion segmentation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0094] In one embodiment, a computer device is provided, whose internal structure diagram can be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a communication interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer-readable instructions and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The database of the computer device is used to store point cloud motion segmentation results, target point cloud motion segmentation results, etc. The communication interface of the computer device is used to connect and communicate with the vehicle-mounted sensor. When the computer-readable instructions are executed by the processor, a point cloud data motion segmentation method is implemented.

[0095] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0096] A computer device includes a memory and one or more processors, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the one or more processors perform the following steps:

[0097] Get multi-frame point cloud data;

[0098] Perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0099] Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0100] Perform cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0101] Perform occlusion analysis on the 3D bounding box to determine the point cloud motion state corresponding to the 3D bounding box;

[0102] When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0103] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: determining a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame in the three-dimensional bounding box; performing occlusion analysis on the current three-dimensional bounding box and the historical three-dimensional bounding box to identify the point cloud motion state corresponding to the current three-dimensional bounding box.

[0104] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: obtaining the point cloud scan lines corresponding to each historical three-dimensional bounding box; identifying the positional relationship between the current three-dimensional bounding box and the point cloud scan lines; and identifying the point cloud motion state corresponding to the current bounding box based on the positional relationship.

[0105] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: extracting dynamic point cloud data corresponding to the three-dimensional bounding box; determining the point cloud motion segmentation result corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation result; when the determined point cloud motion segmentation result is a static point, replacing the determined point cloud motion segmentation result with a dynamic point.

[0106] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: dividing the point cloud area corresponding to each frame of point cloud data into multiple grids; calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculating the distance value between each point in the point cloud data of each grid and the corresponding ground; filtering the points whose distance values ​​are less than a first threshold value to obtain multiple frames of non-ground point cloud data.

[0107] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: selecting the point with the smallest height value in the point cloud data corresponding to each grid; calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; extracting points whose height difference is less than a second threshold, and performing plane fitting on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

[0108] In one of the embodiments, when the processor executes the computer-readable instructions, it also implements the following steps: calling a pre-trained point cloud motion segmentation model and inputting the point cloud data into the point cloud motion segmentation model; classifying the point cloud data through the point cloud motion segmentation model, outputting the motion state corresponding to each point in the point cloud data, and generating a point cloud motion segmentation result.

[0109] One or more non-volatile computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps:

[0110] Get multi-frame point cloud data;

[0111] Perform point cloud motion segmentation on the point cloud data to obtain the point cloud motion segmentation result;

[0112] Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data;

[0113] Perform cluster analysis on multiple frames of non-ground point cloud data to obtain the three-dimensional bounding box corresponding to each frame of non-ground point cloud data;

[0114] Performing occlusion analysis on the three-dimensional bounding box to determine the motion state of the point cloud corresponding to the three-dimensional bounding box; and

[0115] When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result.

[0116] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: determining a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame in the three-dimensional bounding box; performing occlusion analysis on the current three-dimensional bounding box and the historical three-dimensional bounding box to identify the point cloud motion state corresponding to the current three-dimensional bounding box.

[0117] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: obtaining the point cloud scan lines corresponding to each historical three-dimensional bounding box; identifying the positional relationship between the current three-dimensional bounding box and the point cloud scan lines; and identifying the point cloud motion state corresponding to the current bounding box based on the positional relationship.

[0118] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: extracting dynamic point cloud data corresponding to the three-dimensional bounding box; determining the point cloud motion segmentation results corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation results; when the determined point cloud motion segmentation results are static points, replacing the determined point cloud motion segmentation results with dynamic points.

[0119] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: dividing the point cloud area corresponding to each frame of point cloud data into multiple grids; calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculating the distance value between each point in the point cloud data of each grid and the corresponding ground; filtering points with distance values ​​less than a first threshold to obtain multiple frames of non-ground point cloud data.

[0120] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: selecting the point with the smallest height value in the point cloud data corresponding to each grid; calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; extracting points whose height difference is less than a second threshold, and performing plane fitting on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

[0121] In one of the embodiments, when the computer-readable instructions are executed by the processor, the following steps are also implemented: calling a pre-trained point cloud motion segmentation model and inputting the point cloud data into the point cloud motion segmentation model; classifying the point cloud data through the point cloud motion segmentation model, outputting the motion state corresponding to each point in the point cloud data, and generating a point cloud motion segmentation result.

[0122] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through computer-readable instructions, and the computer-readable instructions can be stored in a non-volatile computer-readable storage medium. When the computer-readable instructions are executed, they may include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0123] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0124] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A point cloud data motion segmentation method, comprising: Get multi-frame point cloud data; Performing point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result; Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data; Performing cluster analysis on the multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data; Determine, in the three-dimensional bounding box, a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame; Obtaining point cloud scan lines corresponding to each of the historical three-dimensional bounding boxes; Identifying a positional relationship between the current three-dimensional bounding box and the point cloud scan line; Identify the point cloud motion state corresponding to the current three-dimensional bounding box according to the positional relationship; and When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain a target point cloud motion segmentation result.

2. The method according to claim 1, characterized in that The updating of the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain the target point cloud motion segmentation result includes: Extracting dynamic point cloud data corresponding to the three-dimensional bounding box; Determining the point cloud motion segmentation result corresponding to each dynamic point in the dynamic point cloud data from the point cloud motion segmentation result; and When the determined point cloud motion segmentation result is a static point, the determined point cloud motion segmentation result is replaced with a dynamic point.

3. The method according to claim 1, characterized in that: The performing ground filtering processing on each frame of point cloud data to obtain multiple frames of non-ground point cloud data includes: Divide the point cloud area corresponding to each frame of point cloud data into multiple grids; Calculate the ground corresponding to the point cloud data in each grid according to the preset plane equation; Calculate the distance between each point in the point cloud data of each grid and the corresponding ground; and Points whose distance values ​​are less than the first threshold are filtered to obtain multiple frames of non-ground point cloud data.

4. The method according to claim 3, characterized in that Calculating the ground corresponding to the point cloud data in each grid according to the preset plane equation includes: Select the point with the smallest height value in the point cloud data corresponding to each grid; Calculate the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; and Points whose height difference is less than the second threshold are extracted, and plane fitting is performed on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

5. The method according to any one of claims 1 to 4, characterized in that: The performing point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result comprises: Calling a pre-trained point cloud motion segmentation model, and inputting the point cloud data into the point cloud motion segmentation model; and The point cloud data is classified by the point cloud motion segmentation model, the motion state corresponding to each point in the point cloud data is output, and a point cloud motion segmentation result is generated.

6. A point cloud data motion segmentation device, comprising: Communication module, used to obtain multi-frame point cloud data; A point cloud motion segmentation module is used to perform point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result; A ground filtering module is used to perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data; A cluster analysis module, used to perform cluster analysis on the multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data; The occlusion analysis module is used to determine, in the three-dimensional bounding box, a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame; obtain the point cloud scan lines corresponding to each of the historical three-dimensional bounding boxes; identify the positional relationship between the current three-dimensional bounding box and the point cloud scan lines; and identify the point cloud motion state corresponding to the current three-dimensional bounding box according to the positional relationship; and The updating module is used to update the point cloud motion segmentation result according to the non-ground point cloud data corresponding to the three-dimensional bounding box when the point cloud motion state is dynamic, so as to obtain the target point cloud motion segmentation result.

7. The device according to claim 6, characterized in that The update module is also used to extract the dynamic point cloud data corresponding to the three-dimensional bounding box; determine the point cloud motion segmentation results corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation results; and when the determined point cloud motion segmentation results are static points, replace the determined point cloud motion segmentation results with dynamic points.

8. The device according to claim 6, characterized in that The ground filtering module is also used to divide the point cloud area corresponding to each frame of point cloud data into multiple grids; calculate the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculate the distance value between each point in the point cloud data of each grid and the corresponding ground; and filter the points whose distance value is less than a first threshold to obtain multiple frames of non-ground point cloud data.

9. The device according to claim 8, characterized in that The ground filtering module is also used to select the point with the smallest height value in the point cloud data corresponding to each grid; calculate the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; and extract points whose height difference is less than a second threshold, and perform plane fitting on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

10. The device according to any one of claims 6 to 9, characterized in that The point cloud motion segmentation module is also used to call a pre-trained point cloud motion segmentation model, input the point cloud data into the point cloud motion segmentation model; and classify the point cloud data through the point cloud motion segmentation model, output the motion state corresponding to each point in the point cloud data, and generate a point cloud motion segmentation result.

11. A computer device comprising a memory and one or more processors, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the one or more processors perform the following steps: Get multi-frame point cloud data; Performing point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result; Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data; Performing cluster analysis on the multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data; Determine, in the three-dimensional bounding box, a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame; Obtaining point cloud scan lines corresponding to each of the historical three-dimensional bounding boxes; Identifying a positional relationship between the current three-dimensional bounding box and the point cloud scan line; Identify the point cloud motion state corresponding to the current three-dimensional bounding box according to the positional relationship; and When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain a target point cloud motion segmentation result.

12. The computer device according to claim 11, characterized in that When the processor executes the computer-readable instructions, the processor also performs the following steps: extracting the dynamic point cloud data corresponding to the three-dimensional bounding box; determining the point cloud motion segmentation results corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation results; and when the determined point cloud motion segmentation results are static points, replacing the determined point cloud motion segmentation results with dynamic points.

13. The computer device according to claim 11, characterized in that When the processor executes the computer-readable instructions, the processor also performs the following steps: dividing the point cloud area corresponding to each frame of point cloud data into multiple grids; calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculating the distance value between each point in the point cloud data of each grid and the corresponding ground; and filtering points whose distance values ​​are less than a first threshold to obtain multiple frames of non-ground point cloud data.

14. The computer device according to claim 13, characterized in that When the processor executes the computer-readable instructions, the processor also performs the following steps: selecting the point with the smallest height value in the point cloud data corresponding to each grid; calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; and extracting points whose height difference is less than a second threshold, and performing plane fitting on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

15. The computer device according to any one of claims 11 to 14, characterized in that: When the processor executes the computer-readable instructions, it also performs the following steps: calling a pre-trained point cloud motion segmentation model and inputting the point cloud data into the point cloud motion segmentation model; and classifying the point cloud data through the point cloud motion segmentation model, outputting the motion state corresponding to each point in the point cloud data, and generating a point cloud motion segmentation result.

16. One or more non-volatile computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps: Get multi-frame point cloud data; Performing point cloud motion segmentation on the point cloud data to obtain a point cloud motion segmentation result; Perform ground filtering on each frame of point cloud data to obtain multiple frames of non-ground point cloud data; Performing cluster analysis on the multiple frames of non-ground point cloud data to obtain a three-dimensional bounding box corresponding to each frame of non-ground point cloud data; Determine, in the three-dimensional bounding box, a current three-dimensional bounding box corresponding to the non-ground point cloud data of the current frame and a historical three-dimensional bounding box corresponding to the non-ground point cloud data of the historical frame; Obtaining point cloud scan lines corresponding to each of the historical three-dimensional bounding boxes; Identifying a positional relationship between the current three-dimensional bounding box and the point cloud scan line; Identify the point cloud motion state corresponding to the current three-dimensional bounding box according to the positional relationship; and When the point cloud motion state is dynamic, the point cloud motion segmentation result is updated according to the non-ground point cloud data corresponding to the three-dimensional bounding box to obtain a target point cloud motion segmentation result.

17. The storage medium according to claim 16, characterized in that When the computer-readable instructions are executed by the processor, the following steps are also performed: extracting the dynamic point cloud data corresponding to the three-dimensional bounding box; determining the point cloud motion segmentation results corresponding to each dynamic point in the dynamic point cloud data in the point cloud motion segmentation results; and when the determined point cloud motion segmentation results are static points, replacing the determined point cloud motion segmentation results with dynamic points.

18. The storage medium according to claim 16, characterized in that When the computer-readable instructions are executed by the processor, the following steps are also performed: dividing the point cloud area corresponding to each frame of point cloud data into multiple grids; calculating the ground corresponding to the point cloud data in each grid according to a preset plane equation; calculating the distance value between each point in the point cloud data of each grid and the corresponding ground; and filtering points whose distance values ​​are less than a first threshold to obtain multiple frames of non-ground point cloud data.

19. The storage medium according to claim 18, characterized in that When the computer-readable instructions are executed by the processor, the following steps are also performed: selecting the point with the smallest height value in the point cloud data corresponding to each grid; calculating the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; and extracting points whose height difference is less than a second threshold, and performing plane fitting on the selected points according to a preset plane equation to obtain the ground corresponding to the point cloud data in each grid.

20. The storage medium according to any one of claims 16 to 19, characterized in that: When the computer-readable instructions are executed by the processor, the following steps are also performed: calling a pre-trained point cloud motion segmentation model and inputting the point cloud data into the point cloud motion segmentation model; and classifying the point cloud data through the point cloud motion segmentation model, outputting the motion state corresponding to each point in the point cloud data, and generating a point cloud motion segmentation result.

Images