Obstacle recognition method, device and electronic equipment

Abstract

The present application provides an obstacle recognition method, device and electronic device, the method comprising: first obtaining a three-dimensional point cloud data set of the current position of the vehicle, and dividing the three-dimensional point cloud data set into multiple grids according to the coordinates of each point cloud data, and then grouping the grids with high similarity into a target grid cluster according to the distribution information corresponding to the point cloud data contained in each grid, and finally determining the obstacle information according to the point cloud data contained in the target grid cluster. Compared with the prior art, the present application does not perform ground segmentation on the point cloud data, but directly maps the three-dimensional point cloud to the grid, and clusters the grid according to the distribution information, retaining the distribution characteristics of all three-dimensional point clouds, so that the final obstacle recognition result has better robustness.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to an obstacle recognition method, device and electronic equipment. Background Art

[0002] Environmental perception plays an extremely important role as the eyes of autonomous vehicles. During the autonomous driving process, the vehicle must perceive the surrounding environment in real time to make its own response strategy. LiDAR stands out among many sensors with its high precision. With the widespread popularity of LiDAR in the field of autonomous driving, research on LiDAR point cloud algorithms is in full swing.

[0003] At present, there are two types of lidar obstacle perception: deep learning network algorithm detection and traditional algorithm detection. On the one hand, deep learning network algorithm detection requires a large amount of data collection and annotation. On the other hand, if the scenes are not covered in the annotated data, the deep learning network will not be able to detect obstacles. The first goal of autonomous driving vehicles is to ensure safety. In view of the shortcomings of deep learning network detection algorithms, traditional algorithm detection can achieve the goal of full inspection and ensure driving safety. However, the cumbersome process of traditional algorithm detection and the complexity of the algorithm are currently facing a major challenge. Since autonomous driving has high requirements for system real-time performance, traditional algorithms cannot respond quickly to emergencies, posing a major safety hazard. Summary of the invention

[0004] In view of this, the purpose of this application is to provide an obstacle recognition method, device and electronic equipment to improve the safety of autonomous driving.

[0005] In a first aspect, an embodiment of the present application provides an obstacle identification method, the method comprising: obtaining a three-dimensional point cloud data set of a vehicle's current location; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle; according to the coordinates of each point cloud data in the three-dimensional point cloud data set, the three-dimensional point cloud data set is divided into a grid set consisting of multiple grids; obtaining distribution information corresponding to the point cloud data contained in each grid; wherein the distribution information includes a mean and a covariance matrix of the point cloud data contained in the grid; clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold; determining the obstacle information corresponding to the target grid cluster according to the point cloud data contained in the target grid cluster.

[0006] Furthermore, the above-mentioned step of obtaining a three-dimensional point cloud data set at the current position of the vehicle includes: obtaining an initial point cloud data set; wherein the initial point cloud data set is three-dimensional data obtained by a laser radar sensor installed on the vehicle; converting each initial point cloud data in the initial point cloud set into point cloud data in a vehicle body coordinate system; wherein the origin of the vehicle body coordinate system is the center point of the rear axle of the vehicle, the x-axis is the front direction of the vehicle, the y-axis is the left direction of the vehicle, and the z-axis is the upward direction of the vehicle; and determining the set consisting of the point cloud data in the converted vehicle body coordinate system as a three-dimensional point cloud data set.

[0007] Furthermore, the above-mentioned step of dividing the three-dimensional point cloud data set into a grid set composed of multiple grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set includes: dividing the predetermined spatial range corresponding to the three-dimensional point cloud data set into multiple grids according to preset sizes; determining the grid corresponding to each point cloud data according to the coordinates of each point cloud data in the three-dimensional point cloud data set and the endpoint coordinates of each grid; and determining the grid corresponding to each point cloud data as a grid set.

[0008] Furthermore, the mean value u corresponding to the point cloud data contained in the above grid is determined by the following formula: u=[u x ,u y ,u z ] T ; Where N is the number of point cloud data contained in the grid, x i ,y i and z i The coordinates of each point cloud data.

[0009] Furthermore, the covariance matrix ∑ corresponding to the point cloud data contained in the above grid is determined by the following formula: in, The values ​​of A and B are the coordinates x, y or z of each point cloud data, and N is the number of point cloud data contained in the grid.

[0010] Furthermore, before the above-mentioned step of obtaining distribution information corresponding to the point cloud data contained in each grid, the method also includes: determining a first attribute of the grid based on the amount of point cloud data contained in each grid; wherein the first attribute is used to characterize whether the grid contains point cloud data; and deleting the grids in the grid set whose first attribute characterizes empty grids.

[0011] Furthermore, the above-mentioned first attribute includes empty grids and non-empty grids; the step of determining the first attribute of the grid according to the number of point cloud data contained in each grid includes: judging whether the number of point cloud data contained in the grid is greater than a preset first quantity threshold; if so, determining the first attribute of the grid as a non-empty grid; otherwise, determining the first attribute of the grid as an empty grid.

[0012] Furthermore, before the above-mentioned step of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, the method also includes: solving the covariance matrix of each grid to obtain the eigenvalue set corresponding to the grid; determining the second attribute of the grid according to the eigenvalue set; wherein the second attribute is used to characterize the shape of the object constituted by the point cloud data contained in the grid; and deleting the grids in the grid set whose second attributes represent the ground grid.

[0013] Furthermore, the above-mentioned second attribute includes ground grid and non-ground grid; the step of determining the second attribute of the grid according to the eigenvalue set includes: determining the eigenvectors corresponding to the three eigenvalues ​​with the largest values ​​in the eigenvalue set as the first feature set corresponding to the grid; judging whether the first feature set satisfies the preset ground features; if not, determining that the second attribute of the grid is a non-ground grid; otherwise, determining that the second attribute of the grid is a ground grid.

[0014] Furthermore, the step of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain the target grid cluster includes: determining a second grid adjacent to the first grid according to a preset neighborhood rule; wherein the first grid and the second grid are grids in the grid set; determining a first probability density function of the first grid according to the distribution information of the first grid, and determining a second probability density function of the second grid according to the distribution information of the second grid; determining the similarity between the first grid and the second grid according to the first probability density function and the second probability density function; judging whether the similarity is greater than a grid similarity threshold; if so, marking the third attribute of the first grid and the third attribute of the second grid as a first category; wherein the third attribute is used to characterize the category of the object represented by the point cloud data contained in the grid; determining a set of networks whose third attributes are of the first category as the target grid cluster.

[0015] Furthermore, the above similarity is JS similarity; the similarity between the first grid and the second grid is calculated by the following formula: Wherein, p1 represents the probability density function of the Gaussian distribution of the first grid, p2 represents the probability density function of the Gaussian distribution of the second grid, and KL represents the KL divergence between the first grid and the second grid.

[0016] In a second aspect, an embodiment of the present application also provides an obstacle identification device, which includes: a point cloud data acquisition module, used to acquire a three-dimensional point cloud data set of the current location of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in the vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle; a division module, used to divide the three-dimensional point cloud data set into a grid set consisting of multiple grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set; a distribution information acquisition module, used to acquire distribution information corresponding to the point cloud data contained in each grid; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid; a clustering module, used to cluster the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, to obtain a target grid cluster; wherein the similarity between each two grids in the target grid cluster is greater than the grid similarity threshold; an obstacle determination module, used to determine the obstacle information corresponding to the target grid cluster according to the point cloud data contained in the target grid cluster.

[0017] In a third aspect, an embodiment of the present application further provides an electronic device, including a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the obstacle recognition method of the first aspect above.

[0018] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the obstacle recognition method of the first aspect above.

[0019] Compared with the prior art, this application has the following beneficial effects:

[0020] The obstacle recognition method, device and electronic device provided in the embodiments of the present application first obtain a three-dimensional point cloud data set of the current position of the vehicle, and divide the three-dimensional point cloud data set into multiple grids according to the coordinates of each point cloud data, and then group the grids with high similarity into a target grid cluster according to the distribution information corresponding to the point cloud data contained in each grid, and finally determine the obstacle information according to the point cloud data contained in the target grid cluster. Compared with the prior art, the present application does not perform ground segmentation on the point cloud data, but directly maps the three-dimensional point cloud to the grid, and clusters the grid according to the distribution information, retaining the distribution characteristics of all three-dimensional point clouds, so that the final obstacle recognition result has better robustness.

[0021] Other features and advantages of the present disclosure will be set forth in the following description, or some features and advantages may be inferred or unambiguously determined from the description, or may be learned by implementing the above-mentioned technology of the present disclosure.

[0022] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 A schematic diagram of the structure of an electronic system provided in an embodiment of the present application;

[0025] Figure 2 A flowchart of an obstacle identification method provided in an embodiment of the present application;

[0026] Figure 3 A flowchart of another obstacle identification method provided in an embodiment of the present application;

[0027] Figure 4 A schematic diagram of different shapes corresponding to a three-dimensional Gaussian distribution provided in an embodiment of the present application;

[0028] Figure 5 A grid schematic diagram with grid attributes marked provided in an embodiment of the present application;

[0029] Figure 6 A flowchart of a method for clustering grids provided in an embodiment of the present application;

[0030] Figure 7 A schematic diagram of a point cloud data set labeled with different categories provided in an embodiment of the present application;

[0031] Figure 8 A schematic diagram of the structure of an obstacle recognition device provided in an embodiment of the present application;

[0032] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution of the present application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0034] At present, the target recognition technology in the process of autonomous driving, the general process of traditional algorithm detection is: the first step is to perform a ground segmentation algorithm (RANSAC, based on radial gradient detection), and the second step is to use a clustering method (Euclidean distance clustering, DBSCAN, K-MEANS) or an existing grid method based on the removal of the ground point cloud. The existing grid method directly calculates the mean in the vertical direction of the grid, and then directly clusters according to the relevant threshold of the mean. The disadvantage of the above first step is that the ground segmentation point cloud is not clean, which will lead to false detection in subsequent processing, seriously affecting the driving performance of autonomous driving; the above second step algorithm has the problem of traversing all point clouds, resulting in a relatively high algorithm complexity, and then the existing grid method directly uses the information in the vertical direction of the point cloud, ignoring the horizontal distribution characteristics of the point cloud, and losing a large amount of point cloud information. Based on this, the embodiment of the present application provides an obstacle recognition method, device and electronic device to improve the safety of autonomous driving.

[0035] Reference Figure 1 The electronic system 100 is a schematic diagram of the structure of the electronic system 100. The electronic system can be used to implement the obstacle recognition method and device of the embodiment of the present application.

[0036] like Figure 1 The electronic system 100 includes one or more processing devices 102 and one or more storage devices 104. Optionally, the electronic system 100 may also include an input device 106, an output device 108, and one or more point cloud data acquisition devices 110. These components are interconnected via a bus system 112 and / or other forms of connection mechanisms (not shown). It should be noted that Figure 1 The components and structure of the electronic system 100 shown are only exemplary and not restrictive. The electronic system may have Figure 1 Some components in it may also have other components and structures.

[0037] The processing device 102 can be a server, an intelligent terminal, or a device including a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities. It can process data of other components in the electronic system 100 and can also control other components in the electronic system 100 to perform obstacle recognition functions.

[0038] The storage device 104 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on a computer-readable storage medium, and the processing device 102 may run the program instructions to implement the client functions and / or other desired functions in the embodiments of the present application (implemented by the processing device) described below. Various applications and various data, such as various data used and / or generated by the application, may also be stored in the computer-readable storage medium.

[0039] The input device 106 may be a device used by a user to input instructions, and may include one or more of a keyboard, a mouse, a microphone, a touch screen, and the like.

[0040] The output device 108 may output various information (eg, images or sounds) to the outside (eg, a user), and may include one or more of a display, a speaker, and the like.

[0041] The point cloud data acquisition device 110 can acquire the point cloud data to be processed, and store the point cloud data in the storage device 104 for use by other components.

[0042] Exemplarily, the various devices used to implement the obstacle recognition method, device, and electronic device according to the embodiments of the present application can be integrated or dispersed, such as integrating the processing device 102, the storage device 104, the input device 106, and the output device 108 into one, and setting the point cloud data acquisition device 110 at a specified location where point cloud data can be collected. When the various devices in the above electronic system are integrated, the electronic system can be implemented as a smart terminal such as a camera, a smart phone, a tablet computer, a computer, a vehicle-mounted terminal, etc.

[0043] Figure 2 A flowchart of an obstacle identification method provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the method specifically comprises the following steps:

[0044] S202: Acquire a three-dimensional point cloud data set of the current position of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle;

[0045] Sensors installed on autonomous vehicles, such as radars, lasers, cameras, etc., can acquire three-dimensional point cloud data around the vehicle. At the current moment, an initial point cloud data set is acquired through the sensor, and the coordinate system of the initial point cloud data set is transformed to obtain a set of point cloud data in the vehicle body coordinate system.

[0046] S204: dividing the three-dimensional point cloud data set into a grid set consisting of a plurality of grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set;

[0047] The space corresponding to the three-dimensional point cloud is divided into equal or unequal grids. Each grid may include one or more point cloud data or may not include point cloud data. In this way, a corresponding relationship is established between the point cloud data and the grid.

[0048] S206: Obtaining distribution information corresponding to the point cloud data contained in each grid; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid;

[0049] The mean and covariance matrix in the distribution information are both data that characterize the density information of the point cloud data in the grid. The mean includes the mean in the x direction, the mean in the y direction, and the mean in the z direction, as well as the total mean of the means in the x, y, and z directions. The covariance matrix is ​​calculated by the mean, and the specific calculation method will be explained in detail below, so it will not be repeated here.

[0050] S208: clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold;

[0051] After obtaining the distribution information, the grids with higher similarity are classified into the same category according to the distribution information to obtain one or more grid clusters. The similarity between grids can be JS divergence, KL divergence, or other similarities in the prior art, which is not limited in the embodiments of the present application. Specifically, a grid similarity threshold can be set to classify two grids with a similarity greater than the grid similarity threshold into the same category. After traversing all grids, a grid cluster consisting of all grids of the same category is obtained.

[0052] S210: Determine obstacle information corresponding to the target grid cluster according to the point cloud data included in the target grid cluster.

[0053] After determining the target grid cluster, the point clouds contained in all grids in the grid cluster can be clustered to form a point cloud cluster, and the relevant attributes of the point cloud cluster can be calculated, such as the minimum outer envelope, centroid, center point, etc. According to the relevant attributes of the point cloud cluster, the final obstacle information can be identified.

[0054] The above obstacle recognition method provided by the present application first obtains a set of three-dimensional point cloud data of the current position of the vehicle, and divides the set of three-dimensional point cloud data into multiple grids according to the coordinates of each point cloud data, and then groups the grids with high similarity into target grid clusters according to the distribution information corresponding to the point cloud data contained in each grid, and finally determines the obstacle information according to the point cloud data contained in the target grid cluster. Compared with the prior art, the present application does not perform ground segmentation on the point cloud data, but directly maps the three-dimensional point cloud to the grid, and clusters the grid according to the distribution information, retaining the distribution characteristics of all three-dimensional point clouds, so that the final obstacle recognition result has better robustness.

[0055] In some possible implementations, after obtaining the current initial point cloud data set of the vehicle, the initial point cloud data set needs to be converted into point cloud data in a vehicle body coordinate system. Specifically, the following method can be used for coordinate conversion: obtain an initial point cloud data set; wherein the initial point cloud data set is three-dimensional data obtained by a lidar sensor installed on the vehicle; each initial point cloud data in the initial point cloud set is converted into point cloud data in a vehicle body coordinate system; wherein the origin of the vehicle body coordinate system is the center point of the rear axle of the vehicle, the x-axis is the front direction of the vehicle, the y-axis is the left direction of the vehicle, and the z-axis is the upward direction of the vehicle; the set consisting of the point cloud data in the converted vehicle body coordinate system is determined as a three-dimensional point cloud data set.

[0056] Specifically, the vehicle body coordinate system is: the coordinate system is established by projecting the center of the vehicle's rear axle to the ground, with the front of the vehicle as the x-axis, the left as the y-axis, and the z-axis upward, obeying the right-hand coordinate system. The laser radar and the vehicle body are rigidly connected, and the relative position and posture between the two remain unchanged. The laser radar coordinate system is mapped to the vehicle body coordinate system through the rotation matrix R and translation matrix T parameters. The specific coordinate conversion formula is as follows:

[0057]

[0058] Among them, p′(x′, y′, z′) is the point in the body coordinate system after transformation, p(x, y, z) is the point in the LiDAR coordinate system, R is a 3*3 rotation matrix, and T is a 3*1 translation matrix. The point cloud is converted to the body coordinate system, and all subsequent operations on the point cloud are completed in the body coordinate system.

[0059] In some other possible implementations, the above step S204 (dividing the three-dimensional point cloud data set into a grid set consisting of multiple grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set) can be specifically determined in the following manner:

[0060] (1) dividing a predetermined spatial range corresponding to the three-dimensional point cloud data set into a plurality of grids according to a preset size;

[0061] (2) determining the grid corresponding to each point cloud data according to the coordinates of each point cloud data in the three-dimensional point cloud data set and the endpoint coordinates of each grid;

[0062] (3) The grid corresponding to each point cloud data is determined as a grid set.

[0063] Specifically, the space corresponding to the three-dimensional point cloud data set is divided into grids (two-dimensional) or voxels (three-dimensional) of a specified size. In order to improve the calculation efficiency, the two-dimensional grid method is adopted in the embodiment of the present application to establish a mapping relationship between the two-dimensional plane grid and the point cloud. The vehicle coordinate system xy is used as the plane reference. min , x max ]、[y min ,y max ]Create a grid within the rectangular range, and set the grid size in the x direction to size x , take the y-direction grid size as size y Finally, the grid size is established Calculate the grid number cloNum and rowNum of each point according to the distance.

[0064]

[0065]

[0066] For example, if the grid size is set to 20*10, the length of each grid is 2 and the width is 1, then a grid of 10*10=100 grids can be established.

[0067] The following is a detailed description of how to determine the distribution information corresponding to the point cloud data contained in each grid. The above distribution information includes the mean value corresponding to the point cloud data contained in the grid, and the mean value u can be determined by the following formula:

[0068]

[0069]

[0070]

[0071] u=[u x ,u y ,u z ] T

[0072] Among them, N is the number of point cloud data contained in the grid, x i ,y i and z i The coordinates of each point cloud data.

[0073] The above distribution information includes the covariance matrix corresponding to the point cloud data contained in the grid, and the covariance matrix ∑ is determined by the following formula:

[0074]

[0075] in, The values ​​of A and B are the coordinates x, y or z of each point cloud data, and N is the number of point cloud data contained in the grid.

[0076] In order to make the clustering of grids more efficient, it is necessary to exclude grids without possible obstacles before clustering. Based on this, the embodiment of the present application also provides another obstacle identification method, such as Figure 3 As shown, the method specifically comprises the following steps:

[0077] S302: Acquire a three-dimensional point cloud data set of the current position of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle;

[0078] S304: dividing the three-dimensional point cloud data set into a grid set consisting of a plurality of grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set;

[0079] S306: determining a first attribute of each grid according to the amount of point cloud data contained in the grid; wherein the first attribute is used to characterize whether the grid contains point cloud data;

[0080] S308: deleting the grid whose first attribute represents an empty grid in the grid set;

[0081] The first attribute includes an empty grid and a non-empty grid. Specifically, the first attribute of the grid can be determined by the following method: (1) determining whether the amount of point cloud data contained in the grid is greater than a preset first quantity threshold; (2) if so, determining the first attribute of the grid as a non-empty grid; (3) otherwise, determining the first attribute of the grid as an empty grid.

[0082] For example, the number of point cloud data included in the current grid is N, N=20, and the first quantity threshold is set to 25. Then, the number of point cloud data included in the current grid is less than the first quantity threshold, and the current grid is marked as an empty grid.

[0083] S310: Obtaining distribution information corresponding to the point cloud data contained in each grid; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid;

[0084] S312: Solve the covariance matrix of each grid to obtain the eigenvalue set corresponding to the grid;

[0085] S314: determining a second attribute of the grid according to the eigenvalue set; wherein the second attribute is used to characterize the shape of the object formed by the point cloud data contained in the grid;

[0086] S316: deleting the grid whose second attribute represents the ground grid in the grid set;

[0087] Among them, the second attribute includes ground grids and non-ground grids. The above steps have calculated the covariance matrix of each grid. Then the SVD method can be further used to solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix of each grid, and the first three largest eigenvalues ​​and their corresponding eigenvectors are taken. The different shapes of the three-dimensional Gaussian distribution depend on the relationship between the eigenvalues ​​of the covariance matrix. Figure 4 A schematic diagram of different shapes corresponding to different covariance values ​​in a three-dimensional Gaussian distribution provided in an embodiment of the present application, such as Figure 4 As shown in the figure, the direction of the arrow indicates the eigenvector of the distribution, and the size indicates the eigenvalue. If the three eigenvalues ​​are close, the point cloud data contained in the grid is distributed similarly to a point or a sphere. If one eigenvalue is much larger than the other two eigenvalues, the point cloud data contained in the grid is distributed similarly to a straight line. If one eigenvalue is much smaller than the other two eigenvalues, the normal distribution of the point cloud data contained in the grid describes a plane.

[0088] The distribution characteristics of the grid point cloud are described by the distribution of three-dimensional feature vectors. If the three-dimensional feature vectors conform to the ground features, the grid is marked as a ground grid GRID ground .like Figure 5 As shown, it shows a grid schematic diagram with grid attributes marked provided by an embodiment of the present application, in which the mark '0' is a ground grid and an empty grid, and the mark '1' is an occupied grid.

[0089] Based on the above description, in some possible implementations, the second attribute of the grid can be determined according to the following method: (1) the feature vectors corresponding to the three eigenvalues ​​with the largest values ​​in the eigenvalue set are determined as the first feature set corresponding to the grid; (2) it is determined whether the first feature set satisfies the preset ground features; (3) if not, the second attribute of the grid is determined to be a non-ground grid; (4) otherwise, the second attribute of the grid is determined to be a ground grid.

[0090] S318: clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold;

[0091] S320: Determine obstacle information corresponding to the target grid cluster according to the point cloud data included in the target grid cluster.

[0092] Compared with the existing two-stage lidar point cloud obstacle detection method, the above-mentioned embodiment of the present application adopts a one-stage grid feature extraction method to map the point cloud. After mapping the three-dimensional point cloud to the two-dimensional grid, the point cloud is subsequently indexed by clustering the grid to form a point cloud cluster. Reduce the complexity of the algorithm and improve the efficiency of the algorithm. In addition, by calculating the eigenvalues ​​of the covariance matrix and its eigenvectors to characterize the distribution characteristics of the point cloud in the grid, the three-dimensional point cloud feature information is more fully utilized to distinguish between the ground grid and the occupied grid, which effectively improves the accuracy of obstacle recognition.

[0093] Figure 6 A method for clustering grids provided in an embodiment of the present application is provided. The method performs grid clustering based on a grid set, wherein the grid set may be a grid set obtained after the initial point cloud data division, or an updated grid set in which empty grids are deleted from the grid set, or a grid set in which ground grids are further deleted from the updated grid set, such as Figure 6 As shown, the method specifically comprises the following steps:

[0094] S602: Determine a second grid adjacent to the first grid according to a preset neighborhood rule; wherein the first grid and the second grid are grids in a grid set;

[0095] The preset domain rule may specifically be four neighborhoods or eight neighborhoods. The embodiment of the present application takes eight neighborhoods as an example for explanation:

[0096] (1) First, traverse the grid set and find the grid in the i-th row and j-th column. (i,j) Perform eight-neighborhood traversal and calculate grid (i,j) and grid (i+k,j+k) JS divergence of adjacent grids, where k∈(-1,0,1); set the JS divergence threshold of adjacent grids JS , where grid (i,j) Gaussian distribution is p1, grid (i+k,j+k) The Gaussian distribution is p2.

[0097] (2) Secondly, calculate the JS divergence between p1 and p2. If JS(p1||p2) <Threshold JS , then mark the grid and extract the corresponding grid number for clustering, and then (i+k,j+k) Perform eight-neighborhood traversal and recursively operate. If the marked grid is traversed or JS(p1||p2)>Threshold JS Then it ends.

[0098] (3) If JS(p1||p2)>Threshold JS , then take different k values ​​and calculate the corresponding JS divergence and continue to execute the above step (1).

[0099] (4) Continue to traverse the grids. If the grid has been traversed before, skip it until all grids are traversed.

[0100] S604: determining a first probability density function of the first grid according to the distribution information of the first grid, and determining a second probability density function of the second grid according to the distribution information of the second grid;

[0101] Among them, the probability density function formula of multidimensional Gaussian distribution is as follows:

[0102]

[0103] Where D = 3 dimensions, u and ∑ represent the mean vector and covariance matrix respectively.

[0104] S606: Determine the similarity between the first grid and the second grid according to the first probability density function and the second probability density function;

[0105] The embodiment of the present application measures the similarity of cloud distribution by calculating the JS divergence between Gaussian distributions of adjacent grids. JS divergence is symmetry, and its value is between 0 and 1. If the two distributions are exactly the same, the value is 0; if they are completely different, the value is 1.

[0106] Based on the above description, in some possible implementations, the above similarity is JS similarity; the above similarity is determined by the following formula:

[0107]

[0108] Wherein, p1 represents the probability density function of the Gaussian distribution of the first grid, p2 represents the probability density function of the Gaussian distribution of the second grid, and KL represents the KL divergence between the first grid and the second grid.

[0109] In order to obtain the JS divergence, it is necessary to calculate the KL divergence. The KL divergence in the above JS divergence formula can be calculated by the following formula:

[0110]

[0111] Wherein, d represents the dimension, d=3, and p1 and p2 represent the probability density functions of the Gaussian distribution of adjacent grids.

[0112] S608: Determine whether the similarity is greater than the grid similarity threshold, if yes, execute step S610; otherwise, execute step S612;

[0113] S610: Marking the third attribute of the first grid and the third attribute of the second grid as a first category; wherein the third attribute is used to represent the category of the object represented by the point cloud data contained in the grid;

[0114] S612: marking the third attribute of the first grid and the third attribute of the second grid as different categories;

[0115] S614: Determine whether the similarity of all grids has been calculated, if yes, execute step S616, otherwise, execute step S618;

[0116] S616: Determine a set of all networks whose third attribute is the first category as a target grid cluster;

[0117] S618: Determine the second grid as the first grid, and re-execute step S602.

[0118] The above embodiment of the present application performs eight-domain recursive clustering by calculating the similarity of the probability density functions of adjacent grids, which better reflects the similarity of the distribution of adjacent grid point clouds for clustering. Because the lidar point cloud has three-dimensional spatial properties, the obstacle point cloud and the background point cloud have obvious contrast characteristics. Therefore, the probability density distribution function is used to better distinguish the foreground and background, and at the same time, better characterize the local characteristics of the point cloud.

[0119] Figure 7 A schematic diagram of a point cloud data set with different categories marked in an actual application scenario provided for an embodiment of the present application, wherein the point cloud data in the figure is divided into three categories, namely A, B and C. Furthermore, obstacles are identified for A, B and C respectively to obtain obstacle information.

[0120] Based on the above method embodiment, the present application embodiment also provides an obstacle recognition device, see Figure 8 As shown, the device comprises:

[0121] The point cloud data acquisition module 802 is used to acquire a three-dimensional point cloud data set of the current position of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle;

[0122] A division module 804 is used to divide the three-dimensional point cloud data set into a grid set consisting of a plurality of grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set;

[0123] The distribution information acquisition module 806 is used to acquire the distribution information corresponding to the point cloud data contained in each grid; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid;

[0124] A clustering module 808 is used to cluster the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold;

[0125] The obstacle determination module 810 is used to determine obstacle information corresponding to the target grid cluster according to the point cloud data contained in the target grid cluster.

[0126] The obstacle recognition device provided by the present application first obtains a set of three-dimensional point cloud data of the current position of the vehicle, and divides the set of three-dimensional point cloud data into multiple grids according to the coordinates of each point cloud data, and then groups the grids with high similarity into a target grid cluster according to the distribution information corresponding to the point cloud data contained in each grid, and finally determines the obstacle information according to the point cloud data contained in the target grid cluster. Compared with the prior art, the present application does not perform ground segmentation on the point cloud data, but directly maps the three-dimensional point cloud to the grid, and clusters the grid according to the distribution information, retaining the distribution characteristics of all three-dimensional point clouds, so that the final obstacle recognition result has better robustness.

[0127] The above-mentioned point cloud data acquisition module 802 is also used to: obtain an initial point cloud data set; wherein the initial point cloud data set is three-dimensional data obtained by a laser radar sensor installed on a vehicle; convert each initial point cloud data in the initial point cloud set into point cloud data in a vehicle body coordinate system; wherein the origin of the vehicle body coordinate system is the center point of the rear axle of the vehicle, the x-axis is the front direction of the vehicle, the y-axis is the left direction of the vehicle, and the z-axis is the upward direction of the vehicle; and determine the set of point cloud data in the converted vehicle body coordinate system as a three-dimensional point cloud data set.

[0128] The above-mentioned division module 804 is also used to: divide the predetermined spatial range corresponding to the three-dimensional point cloud data set into multiple grids according to preset sizes; determine the grid corresponding to each point cloud data according to the coordinates of each point cloud data in the three-dimensional point cloud data set and the endpoint coordinates of each grid; and determine the grid corresponding to each point cloud data as a grid set.

[0129] The mean u corresponding to the point cloud data contained in the above grid is determined by the following formula: u=[u x ,u y ,u z ] T ; Where N is the number of point cloud data contained in the grid, xi ,y i and z i The coordinates of each point cloud data.

[0130] The covariance matrix ∑ corresponding to the point cloud data contained in the above grid is determined by the following formula: in, The values ​​of A and B are the coordinates x, y or z of each point cloud data, and N is the number of point cloud data contained in the grid.

[0131] The above-mentioned device also includes: a first attribute determination module, which is used to determine the first attribute of the grid according to the amount of point cloud data contained in each grid; wherein the first attribute is used to characterize whether the grid contains point cloud data; and a first deletion module, which is used to delete the grids in the grid set whose first attribute characterizes empty grids.

[0132] In the above, the first attribute includes an empty grid and a non-empty grid; the above process of determining the first attribute of the grid according to the number of point cloud data contained in each grid includes: determining whether the number of point cloud data contained in the grid is greater than a preset first quantity threshold; if so, determining the first attribute of the grid as a non-empty grid; otherwise, determining the first attribute of the grid as an empty grid.

[0133] The above-mentioned device also includes: an eigenvalue set determination module, which is used to solve the covariance matrix of each grid to obtain the eigenvalue set corresponding to the grid; a second attribute determination module, which is used to determine the second attribute of the grid based on the eigenvalue set; wherein the second attribute is used to characterize the shape of the object composed of the point cloud data contained in the grid; and a second deletion module, which is used to delete the grid whose second attribute represents the ground grid in the grid set.

[0134] The above-mentioned second attributes include ground grids and non-ground grids; the above-mentioned process of determining the second attribute of the grid based on the eigenvalue set includes: determining the eigenvectors corresponding to the three eigenvalues ​​with the largest values ​​in the eigenvalue set as the first feature set corresponding to the grid; judging whether the first feature set satisfies the preset ground features; if not, determining that the second attribute of the grid is a non-ground grid; otherwise, determining that the second attribute of the grid is a ground grid.

[0135] The above process of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain the target grid cluster includes: determining the second grid adjacent to the first grid according to the preset neighborhood rule; wherein the first grid and the second grid are grids in the grid set; determining the first probability density function of the first grid according to the distribution information of the first grid, and determining the second probability density function of the second grid according to the distribution information of the second grid; determining the similarity between the first grid and the second grid according to the first probability density function and the second probability density function; judging whether the similarity is greater than the grid similarity threshold; if so, marking the third attribute of the first grid and the third attribute of the second grid as the first category; wherein the third attribute is used to characterize the category of the object represented by the point cloud data contained in the grid; determining the set consisting of all networks whose third attributes are the first category as the target grid cluster.

[0136] The above similarity is JS similarity; the similarity between the first grid and the second grid is calculated by the following formula: Wherein, p1 represents the probability density function of the Gaussian distribution of the first grid, p2 represents the probability density function of the Gaussian distribution of the second grid, and KL represents the KL divergence between the first grid and the second grid.

[0137] The obstacle recognition device provided in the embodiment of the present application has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the embodiment of the aforementioned device, reference may be made to the corresponding contents in the aforementioned obstacle recognition method embodiment.

[0138] The present application also provides an electronic device, such as Fig. 9 As shown, it is a schematic diagram of the structure of the electronic device, wherein the electronic device includes a processor 1501 and a memory 1502, the memory 1502 stores computer executable instructions that can be executed by the processor 1501, and the processor 1501 executes the computer executable instructions to implement the above-mentioned obstacle recognition method.

[0139] exist Fig. 9 In the illustrated embodiment, the electronic device further includes a bus 1503 and a communication interface 1504 , wherein the processor 1501 , the communication interface 1504 and the memory 1502 are connected via the bus 1503 .

[0140] Among them, the memory 1502 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 1504 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 1503 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 1503 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0141] Processor 1501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 1501. The above processor 1501 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and performed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor 1501 reads the information in the memory and completes the steps of the obstacle recognition method of the above embodiment in combination with its hardware.

[0142] The embodiment of the present application also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by the processor, the computer-executable instructions prompt the processor to implement the above-mentioned obstacle recognition method. The specific implementation can be found in the aforementioned method embodiment, which will not be repeated here.

[0143] The computer program product of the obstacle identification method, device and electronic device provided in the embodiments of the present application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be found in the method embodiments and will not be repeated here.

[0144] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present application.

[0145] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0146] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0147] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An obstacle recognition method, characterized in that: The method comprises: Acquire a three-dimensional point cloud data set of the current position of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle; According to the coordinates of each point cloud data in the three-dimensional point cloud data set, the three-dimensional point cloud data set is divided into a grid set consisting of a plurality of grids; Obtaining distribution information corresponding to the point cloud data contained in each of the grids; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid; Clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold; Determine obstacle information corresponding to the target grid cluster according to the point cloud data contained in the target grid cluster; Before the step of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, the method further comprises: solving the covariance matrix of each grid to obtain the eigenvalue set corresponding to the grid; determining the second attribute of the grid according to the eigenvalue set; wherein the second attribute is used to characterize the shape of the object constituted by the point cloud data contained in the grid; deleting the grid whose second attribute characterizes the ground grid in the grid set; Wherein, the second attribute includes a ground grid and a non-ground grid; the step of determining the second attribute of the grid according to the eigenvalue set includes: determining the eigenvectors corresponding to the three eigenvalues ​​with the largest values ​​in the eigenvalue set as the first feature set corresponding to the grid; judging whether the first feature set satisfies the preset ground features, wherein if any eigenvalue is much larger than the other two eigenvalues, the point cloud data distribution contained in the grid is presented as a straight line, and if any eigenvalue is much smaller than the other two eigenvalues, the point cloud data distribution contained in the grid is presented as a plane; if not, determining that the second attribute of the grid is a non-ground grid; otherwise, determining that the second attribute of the grid is a ground grid.

2. The method according to claim 1, characterized in that The steps of obtaining a three-dimensional point cloud data set of the current position of the vehicle include: Acquire an initial point cloud data set; wherein the initial point cloud data set is three-dimensional data acquired by a laser radar sensor installed on a vehicle; Convert each initial point cloud data in the initial point cloud set into point cloud data in a vehicle body coordinate system; wherein the origin of the vehicle body coordinate system is the center point of the rear axle of the vehicle, the x-axis is the front direction of the vehicle, the y-axis is the left direction of the vehicle, and the z-axis is the upward direction of the vehicle; The set of point cloud data in the converted vehicle body coordinate system is determined as a three-dimensional point cloud data set.

3. The method according to claim 1, characterized in that The step of dividing the three-dimensional point cloud data set into a grid set consisting of a plurality of grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set comprises: Dividing a predetermined spatial range corresponding to the three-dimensional point cloud data set into a plurality of grids according to preset sizes; Determine the grid corresponding to each point cloud data according to the coordinates of each point cloud data in the three-dimensional point cloud data set and the endpoint coordinates of each grid; The grids corresponding to each of the point cloud data are determined as a grid set.

4. The method according to any one of claims 1 to 3, characterized in that: The mean value u corresponding to the point cloud data contained in the grid is determined by the following formula: in=[in x ,in y ,in z ] T Among them, N is the number of point cloud data contained in the grid, x i ,y i and z i The coordinates of each point cloud data.

5. The method according to any one of claims 1 to 3, characterized in that: The covariance matrix ∑ corresponding to the point cloud data contained in the grid is determined by the following formula: in, The values ​​of A and B are the coordinates x, y or z of each point cloud data, and N is the number of point cloud data contained in the grid.

6. The method according to claim 1, characterized in that Before the step of obtaining distribution information corresponding to the point cloud data contained in each of the grids, the method further includes: Determine a first attribute of each grid according to the amount of point cloud data contained in the grid; wherein the first attribute is used to characterize whether the grid contains point cloud data; Delete the grids in the grid set whose first attribute represents an empty grid.

7. The method according to claim 6, characterized in that The first attribute includes an empty grid and a non-empty grid; The step of determining the first attribute of each grid according to the amount of point cloud data contained in the grid comprises: Determining whether the amount of point cloud data contained in the grid is greater than a preset first amount threshold; If yes, determine the first attribute of the grid as a non-empty grid; Otherwise, the first attribute of the grid is determined to be an empty grid.

8. The method according to claim 1, characterized in that The step of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid to obtain a target grid cluster comprises: Determine a second grid adjacent to the first grid according to a preset neighborhood rule; wherein the first grid and the second grid are grids in the grid set; Determine a first probability density function of the first grid according to the distribution information of the first grid, and determine a second probability density function of the second grid according to the distribution information of the second grid; Determine the similarity between the first grid and the second grid according to the first probability density function and the second probability density function; Determining whether the similarity is greater than a grid similarity threshold; If yes, marking the third attribute of the first grid and the third attribute of the second grid as a first category; wherein the third attribute is used to characterize the category of the object represented by the point cloud data contained in the grid; A set of all networks whose third attribute is the first category is determined as a target grid cluster.

9. The method according to claim 8, characterized in that The similarity is JS similarity; The similarity between the first grid and the second grid is calculated by the following formula: Wherein, p1 represents the probability density function of the Gaussian distribution of the first grid, p2 represents the probability density function of the Gaussian distribution of the second grid, and KL represents the KL divergence between the first grid and the second grid.

10. An obstacle recognition device, characterized in that: The device comprises: A point cloud data acquisition module, used to acquire a three-dimensional point cloud data set of the current position of the vehicle; wherein the three-dimensional point cloud data set is a set of point cloud data in a vehicle body coordinate system corresponding to the vehicle; the point cloud data is determined by a laser radar sensor installed on the vehicle; A division module, used for dividing the three-dimensional point cloud data set into a grid set consisting of a plurality of grids according to the coordinates of each point cloud data in the three-dimensional point cloud data set; A distribution information acquisition module, used to acquire distribution information corresponding to the point cloud data contained in each of the grids; wherein the distribution information includes the mean and covariance matrix of the point cloud data contained in the grid; A clustering module, used for clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, to obtain a target grid cluster; wherein the similarity between every two grids in the target grid cluster is greater than a grid similarity threshold; An obstacle determination module, configured to determine obstacle information corresponding to the target grid cluster according to the point cloud data contained in the target grid cluster; Before the step of clustering the grids obtained by dividing the three-dimensional point cloud data set according to the distribution information of each grid, the method further comprises: solving the covariance matrix of each grid to obtain the eigenvalue set corresponding to the grid; determining the second attribute of the grid according to the eigenvalue set; wherein the second attribute is used to characterize the shape of the object constituted by the point cloud data contained in the grid; deleting the grid whose second attribute characterizes the ground grid in the grid set; Among them, the second attribute includes ground grid and non-ground grid; the step of determining the second attribute of the grid according to the eigenvalue set includes: determining the eigenvectors corresponding to the three eigenvalues ​​with the largest values ​​in the eigenvalue set as the first feature set corresponding to the grid; judging whether the first feature set satisfies the preset ground features, wherein if any eigenvalue is much smaller than the other two eigenvalues, the point cloud data contained in the grid is distributed as a plane; if not, determining that the second attribute of the grid is a non-ground grid; otherwise, determining that the second attribute of the grid is a ground grid.

11. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 9.

Images