Radar calibration and verification method, device, computer equipment and storage medium

Abstract

The present application relates to a radar calibration method, apparatus, computer equipment, storage medium, and computer program product. The method includes: obtaining first point cloud data received by a reference radar during its movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; in the process of the reference radar moving to the area where the radar to be calibrated is located, there is at least one position so that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both include point cloud data reflected by the reflectors; the first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain the external parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin. This method can simultaneously complete the calibration of multiple radars to be calibrated, greatly improving calibration efficiency.

Description

Technical Field

[0001] The present application relates to the field of laser radar technology, and in particular to a radar calibration and verification method, apparatus, computer equipment, storage medium, and computer program product. Background Art

[0002] LiDAR detects a target's position and velocity by emitting a laser beam toward it and receiving the reflected beam. LiDAR can provide reliable target information for autonomous driving. To enhance the perception and range of an autonomous vehicle's surroundings, multiple LiDARs are typically installed per vehicle. By combining the fields of view of multiple LiDARs simultaneously, a larger field of view can be achieved. Multiple LiDARs also improve the robustness of the entire system. However, due to different installation locations, the coordinate systems of multiple LiDARs are not uniform. This results in the point clouds output by multiple LiDARs not being unified into the same coordinate system. Therefore, extrinsic calibration of multiple LiDARs is crucial.

[0003] In related technologies, the external parameter calibration of lidar can usually only calibrate one radar at a time, resulting in low calibration efficiency. Summary of the Invention

[0004] Based on this, it is necessary to provide a radar calibration method, device, computer equipment, computer-readable storage medium and computer program product that can improve calibration efficiency in response to the above technical problems.

[0005] In a first aspect, the present application provides a radar calibration method. The method comprises:

[0006] Acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors;

[0007] The first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with an initial position of the reference radar as an origin.

[0008] In one embodiment, after establishing a reference coordinate system with the initial position of the reference radar as the origin, the method further includes:

[0009] Obtaining the initial coordinates of the radar to be calibrated in the reference coordinate system;

[0010] Using the initial coordinates of the radar to be calibrated as initialization matrix parameters of the matching algorithm;

[0011] The matching of the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system includes:

[0012] The first point cloud data and the second point cloud data are matched using a matching algorithm with the initialization matrix parameters to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

[0013] In one embodiment, the matching algorithm with the initialization matrix parameters is used to match the first point cloud data and the second point cloud data to obtain the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system, including:

[0014] generating map data to be matched based on the first point cloud data;

[0015] Obtaining standard normal distribution parameters of the first point cloud data according to the map data to be matched;

[0016] Obtaining standard normal distribution parameters of the second point cloud data according to the second point cloud data and the initialization matrix parameters;

[0017] An extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system is obtained according to the standard normal distribution parameters of the first point cloud data and the standard normal distribution parameters of the second point cloud data.

[0018] In one embodiment, matching the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system includes:

[0019] Acquire first sub-point cloud data in the first point cloud data, the acquisition time range of which is the same as the acquisition time range of the second point cloud data;

[0020] The first sub-point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

[0021] In one embodiment, the reflective plate includes a circular reflective plate, a triangular reflective plate and a polygonal reflective plate.

[0022] In a second aspect, the present application further provides a radar calibration device. The device comprises:

[0023] a point cloud acquisition module configured to acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; wherein, during movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors;

[0024] a point cloud matching module, configured to match the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, wherein the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin.

[0025] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are performed:

[0026] Acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors;

[0027] The first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with an initial position of the reference radar as an origin.

[0028] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0029] Acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors;

[0030] The first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with an initial position of the reference radar as an origin.

[0031] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the following steps:

[0032] Acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors;

[0033] The first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with an initial position of the reference radar as an origin.

[0034] The radar calibration method, apparatus, computer device, storage medium, and computer program product described above obtain first point cloud data received by a reference radar during its movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scene; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both contain point cloud data reflected by the reflectors; and the first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin. This application matches the first point cloud data of the entire scene received by the reference radar during its movement along a preset route with the second point cloud data of the local scene received by multiple radars to be calibrated in the same scene, to obtain the extrinsic parameter matrix of each radar to be calibrated, thereby achieving simultaneous calibration of multiple radars to be calibrated and greatly improving calibration efficiency. At the same time, reflectors of different shapes are installed within the visual range of different radars to be calibrated, which can make the received point cloud data stronger and the calibration results more accurate.

[0035] In a sixth aspect, the present application also provides a radar calibration verification method. The method comprises:

[0036] Acquire third point cloud data received by the radar scanning preset reference plane to be verified;

[0037] Converting the third point cloud data received by each of the radars to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0038] Whether the radar calibration result is accurate is determined based on the flatness of the fourth point cloud data.

[0039] In a seventh aspect, the present application further provides a radar calibration verification device. The device comprises:

[0040] A first verification module is used to obtain third point cloud data received by the radar to be verified scanning the preset reference plane;

[0041] A second verification module is configured to convert the third point cloud data received by each of the radars to be verified into a same coordinate system according to an extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0042] The third verification module is used to determine whether the radar calibration result is accurate based on the flatness of the fourth point cloud data.

[0043] In an eighth aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0044] Acquire third point cloud data received by the radar scanning preset reference plane to be verified;

[0045] Converting the third point cloud data received by each of the radars to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0046] Whether the radar calibration result is accurate is determined based on the flatness of the fourth point cloud data.

[0047] In a ninth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0048] Acquire third point cloud data received by the radar scanning preset reference plane to be verified;

[0049] Converting the third point cloud data received by each of the radars to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0050] Whether the radar calibration result is accurate is determined based on the flatness of the fourth point cloud data.

[0051] In a tenth aspect, the present application further provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the following steps:

[0052] Acquire third point cloud data received by the radar scanning preset reference plane to be verified;

[0053] Converting the third point cloud data received by each of the radars to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0054] Whether the radar calibration result is accurate is determined based on the flatness of the fourth point cloud data.

[0055] The verification method, device, computer equipment, storage medium and computer program product of the above-mentioned radar calibration obtain the third point cloud data received by the radar to be verified when scanning the preset reference plane; convert the third point cloud data received by each of the radars to be verified into the same coordinate system according to the external parameter matrix of the radar to be verified, to obtain the fourth point cloud data; the external parameter matrix is obtained according to the above-mentioned radar calibration method; and the flatness of the fourth point cloud data is used to determine whether the radar calibration result is accurate. The present application converts the third point cloud data received by each radar to be verified into the same coordinate system through the external parameter matrix of the radar to be verified obtained according to the above-mentioned radar calibration method, and judges the consistency between the radars to be verified according to the flatness of the point cloud data in the same coordinate system, thereby verifying whether the radar calibration result is accurate, and realizing accurate verification of the radar to be verified. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A diagram illustrating an application environment of a radar calibration method according to an embodiment;

[0057] Figure 2 1 is a flow chart of a radar calibration method according to an embodiment;

[0058] Figure 3 FIG1 is a flow chart of obtaining an extrinsic parameter matrix of a radar to be calibrated in a reference coordinate system in one embodiment;

[0059] Figure 4 204 is a flow chart of step 204 in one embodiment;

[0060] Figure 5 1 is a flow chart of a radar calibration verification method according to an embodiment;

[0061] Figure 6 1 is a flow chart of a radar calibration and verification method according to an embodiment;

[0062] Figure 7 is a structural block diagram of a radar calibration device in one embodiment;

[0063] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0065] The radar calibration method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the reference radar 104 is installed or placed on an unmanned forklift 102. The unmanned forklift 102 can move along a preset route 108. The preset route 108 passes through the area where the radar to be calibrated 106 is located and the visual range of the radar to be calibrated 106. Reflectors of different shapes (not shown) are installed in the visual range of the radar to be calibrated 106. The unmanned forklift 102 can also be replaced with other autonomous vehicles. The actual application scenario can be large-scale environments such as factories or industrial parks.

[0066] The computer device obtains first point cloud data received by the reference radar during its movement along a preset route, and second point cloud data received by the radar to be calibrated in the same scenario; wherein, during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both contain point cloud data reflected by the reflectors; the computer device matches the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin. The computer device can be a terminal or a server, or the terminal can obtain the first point cloud data and the second point cloud data and send them to the server for processing.

[0067] In one embodiment, Figure 2 As shown, a radar calibration method is provided, which is described by taking the application of the method to a server as an example, and includes the following steps:

[0068] Step 202: Acquire first point cloud data received by a reference radar while it moves along a preset route, and second point cloud data received by the radar to be calibrated in the same scenario. In this case, there is at least one position where the visual range of the reference radar includes the visual range of the radar to be calibrated during the movement of the reference radar to the area where the radar to be calibrated is located. Reflectors of different shapes are installed within the visual ranges of different radars to be calibrated. Both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors.

[0069] The server obtains first point cloud data received by the reference radar while moving along a preset route, as well as second point cloud data received by the radar to be calibrated in the same scene. The same scene refers to having the same visible range area, which can be an indoor scene or an outdoor scene, such as a scene within the same factory or industrial park. In this embodiment, there is only one reference radar and multiple radars to be calibrated, and the locations of the multiple radars to be calibrated are different, that is, the visible range areas of the radars to be calibrated are also different. During the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position where the visible range of the reference radar includes the visible range of the radar to be calibrated. For example, if the preset route passes through the area where the radar to be calibrated is located and the visible range area of the radar to be calibrated, the visible range of the reference radar can include the visible range of the radar to be calibrated. The purpose is to enable the reference radar to scan the point cloud data scanned by the radar to be calibrated, and this is the case for each radar to be calibrated. At the same time, reflectors of different shapes are installed within the visual ranges of different radars to be calibrated. For example, a circular reflector is installed within the visual range of radar A to be calibrated, a triangular reflector is installed within the visual range of radar B to be calibrated, a square reflector is installed within the visual range of radar C to be calibrated, a pentagonal reflector is installed within the visual range of radar D to be calibrated, and so on. Because reflectors of different shapes are installed within the visual ranges of different radars to be calibrated, the radar to be calibrated can receive point cloud data reflected by the reflectors installed within its visual range. However, during the movement of the reference radar, there is a situation where the visual range of the reference radar includes the visual range of the radar to be calibrated. That is, the reference radar can receive point cloud data emitted by the reflectors installed within the visual range received by the radar to be calibrated. That is, the first point cloud data and the second point cloud data both include point cloud data reflected by the reflectors.

[0070] In this embodiment, the different radars to be calibrated are fixedly installed in different locations. Therefore, their visual ranges may or may not overlap. The reference radar acquires first point cloud data along a preset route, while the radars to be calibrated acquire second point cloud data within their corresponding visual ranges. This means that the first point cloud data includes the second point cloud data received by all the radars to be calibrated.

[0071] Step 204 , matching the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, wherein the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin.

[0072] The server matches the first point cloud data and the second point cloud data according to the matching algorithm to obtain the external parameter matrix of the radar to be calibrated in the reference coordinate system. The matching algorithm refers to the point cloud matching algorithm, the purpose of which is to compare the differences between the two and obtain the relationship between the two. Commonly used point cloud matching algorithms include the ICP (Iterative Closest Point) algorithm and the NDT (Normal Distribution Transform) algorithm. The reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin. For example, the initial position of the reference radar can be used as the origin. The initial position of the reference radar can be selected to be within 3-5 meters of any radar to be calibrated, and any two coordinate axes in the reference coordinate system, such as the X-axis and the Y-axis, are constructed on the horizontal plane where the reference radar is located. At the same time, the radar to be calibrated and the reference radar are placed on the same horizontal plane, which can reduce the computational complexity of the matching process.

[0073] The above radar calibration method obtains the first point cloud data received by the reference radar during its movement along a preset route, and the second point cloud data received by the radar to be calibrated in the same scene; in the process of the reference radar moving to the area where the radar to be calibrated is located, there is at least one position so that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed in the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both contain point cloud data reflected by the reflectors; the first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system, which is a coordinate system established with the initial position of the reference radar as the origin. The embodiment of the present application matches the first point cloud data of the entire scene received by the reference radar along the preset route with the second point cloud data of the local scene received by multiple radars to be calibrated in the same scene, and obtains the extrinsic parameter matrix of each radar to be calibrated respectively, so as to realize the calibration of multiple radars to be calibrated at the same time, thereby greatly improving the calibration efficiency. At the same time, reflectors of different shapes are installed within the visual range of different radars to be calibrated, which can make the received point cloud data stronger and the calibration results more accurate.

[0074] In one embodiment, after establishing a reference coordinate system with the initial position of the reference radar as the origin, the method further includes:

[0075] Get the initial coordinates of the radar to be calibrated in the reference coordinate system.

[0076] After the reference coordinate system is established, the offset of each radar to be calibrated relative to the origin in the reference coordinate system can be obtained by manual measurement or equipment measurement, and the initial coordinates of each radar to be calibrated can be obtained.

[0077] The initial coordinates of the radar to be calibrated are used as the initialization matrix parameters of the matching algorithm.

[0078] The initial coordinates of the radar to be calibrated are used as the initial distance value, and the initial distance value is input into the matching algorithm as the initialization extrinsic parameter matrix parameter.

[0079] The first point cloud data and the second point cloud data are matched according to the matching algorithm to obtain the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system, including:

[0080] The first point cloud data and the second point cloud data are matched using a matching algorithm with initialized matrix parameters to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

[0081] This embodiment uses the obtained initial coordinates of the radar to be calibrated as the initialization matrix parameters of the matching algorithm, which is beneficial to improving the matching accuracy of the matching algorithm and can also improve the operation speed of the matching algorithm.

[0082] In one embodiment, Figure 3 As shown, a matching algorithm with initialized matrix parameters is used to match the first point cloud data and the second point cloud data to obtain the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system, including:

[0083] Step 302: Generate map data to be matched based on the first point cloud data.

[0084] The first point cloud data received by the reference radar while moving along the preset route is used to generate map data to be matched using a mapping tool. For example, SLAM (simultaneous localization and mapping) can be used to generate map data to be matched.

[0085] Step 304: Obtain standard normal distribution parameters of the first point cloud data based on the map data to be matched.

[0086] In this embodiment, the map data to be matched can be downsampled using a voxel gridding method to obtain the first target point cloud data, and the standard normal distribution of the first target point cloud data in each grid is calculated to obtain the standard normal distribution of the first point cloud data in each grid. The size and number of grids can be set as needed.

[0087] Step 306 : Obtain standard normal distribution parameters of the second point cloud data according to the second point cloud data and the initialization matrix parameters.

[0088] In this embodiment, the product of the second point cloud data and the initialization matrix parameters can be used as the second point cloud initial data, and the second point cloud initial data can be downsampled according to the voxel gridding method to obtain the second point cloud target data. The probability of each second point cloud target data falling into the grid corresponding to the first point cloud data is calculated to obtain the standard normal distribution parameters corresponding to the second point cloud target data.

[0089] Step 308 : Obtain an extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system according to the standard normal distribution parameters of the first point cloud data and the standard normal distribution parameters of the second point cloud data.

[0090] The first and second point cloud data are downsampled separately according to a voxel gridding method; the space containing the first point cloud data is divided into multiple three-dimensional grids, and for each grid, a corresponding probability density function is calculated based on the point cloud distribution within the grid; for each second point cloud data, each second point cloud data is mapped to the coordinate system containing the first point cloud data according to the initialization matrix parameters to obtain the corresponding mapping point; the probability of each mapping point falling within the corresponding grid is calculated based on the normal distribution parameters of the first point cloud data in the grid, and the score value corresponding to the coordinate transformation parameter is obtained based on the probability; the score value is continuously optimized until a preset convergence condition is met, and the coordinate transformation parameters corresponding to the optimal score value are obtained, i.e., the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system. The preset convergence condition can be a preset number of iterations, a preset score value, etc. For example, the preset convergence condition can also be that when the score value reaches a certain maximum value, and the score values obtained after a preset number of iterations are all less than the maximum value, the iteration is stopped, and the coordinate transformation parameters corresponding to the maximum value are used as the extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system.

[0091] In a specific example, the NDT algorithm is used to match the first point cloud data with the second point cloud data as follows:

[0092] (1) The first point cloud data and the second point cloud data are downsampled separately according to the voxel gridding method. According to the voxel gridding method, the input point cloud data is created as a three-dimensional voxel grid. The voxel grid can be imagined as a collection of tiny three-dimensional cubes. Then, within each voxel, that is, within the three-dimensional cube, the center of gravity of all points in the voxel is used to approximate the other points in the voxel. In this way, all points in the voxel are finally represented by a center of gravity point. After processing all voxels, a filtered point cloud is obtained. That is, while reducing the amount of point cloud data, the shape characteristics of the point cloud are maintained.

[0093] (2) Divide the space where the first point cloud data is located into multiple three-dimensional grids, and for each grid, calculate its probability density function based on the point cloud distribution within the grid.

[0094] Mean:

[0095] in, Represents all the first point cloud data in a grid.

[0096] Covariance matrix:

[0097] The probability density function of a grid is:

[0098] (3) For each second point cloud data, map each second point cloud data to the coordinate system of the first point cloud data according to the initialization matrix parameters to obtain the corresponding mapping point; calculate the probability of each mapping point falling in the corresponding grid according to the normal distribution parameters of the grid The sum of the probabilities of each mapping point falling in the corresponding grid is used as the fractional value of the coordinate transformation parameter T in this round. Conduct an assessment.

[0099]

[0100]

[0101] in, Represents the mapping point, n represents the number of grids corresponding to the mapping point, d1 and d2 represent the constants of the standard normal distribution to the mixed normal distribution, is the mean vector of the mapping points, ∑ k is the mapping point covariance.

[0102] Three-dimensional transformation matrix in NDT algorithm It can be expressed as:

[0103]

[0104] Where, t=[t x t y t z ],r=[r x r y r z ], s=sinΦ, c=cosΦ, t x , t y , t z Represents the position offset on the x, y, and z coordinate axes, r x , r y , r zThey represent the angular offset in the x, y, and z directions respectively, and Φ is the angle between the mapping point and the first point cloud.

[0105] (4) Use Newton optimization algorithm to optimize the above fractional values To optimize, take The Newton algorithm is also called the rapid descent method, and its basic formula is as follows:

[0106] HΔp=-g

[0107] g is the Jacobian matrix, which is expressed as follows:

[0108]

[0109] Indicates the deviation of the mapped point from the mean of the mapped points.

[0110] H is the Hessian matrix, the formula is as follows:

[0111]

[0112] (5) Repeat steps (3) to (4) until the preset convergence condition is met.

[0113] In one embodiment, Figure 4 As shown, the first point cloud data and the second point cloud data are matched according to the matching algorithm to obtain the external parameter matrix of the radar to be calibrated in the reference coordinate system, including:

[0114] Step 402 : Acquire first sub-point cloud data in the first point cloud data, the acquisition time range of which is the same as the acquisition time range of the second point cloud data.

[0115] Since the first point cloud data is the point cloud data received by the reference radar while it moves along a preset route, that is, the first point cloud data includes point cloud data within the visual range of the entire preset route, the first sub-point cloud data within the first point cloud data that falls within the same acquisition time range as the second point cloud data can be selected based on the acquisition time range. In other words, ensuring that the visual range corresponding to the first sub-point cloud data is the same as the visual range corresponding to the second point cloud data can improve the matching degree between the first sub-point cloud data and the second point cloud data. Each second point cloud data received by the radar to be calibrated corresponds to a first sub-point cloud data.

[0116] Step 404 : Match the first sub-point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system.

[0117] The server can match the second point cloud data and the first sub-point cloud data within the corresponding acquisition time range according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated corresponding to the second point cloud data in the reference coordinate system.

[0118] In one embodiment, the reflective sheet includes a circular reflective sheet, a triangular reflective sheet, and a polygonal reflective sheet.

[0119] In this embodiment, reflectors of different shapes are installed within the visual ranges of different radars to be calibrated. The reflectors include circular reflectors, triangular reflectors, quadrilateral reflectors, pentagonal reflectors, and other polygonal reflectors. The specific shape is not limited within the visual range of the same radar to be calibrated; it is only necessary that the visual ranges of different radars to be calibrated correspond to reflectors of different shapes. The size and material of the specific reflectors are selected based on the actual application scenario and are not further defined here. Installing reflectors of different shapes within the visual ranges of different radars to be calibrated can increase the intensity difference between the second point cloud data received by the different radars to be calibrated, and can improve the accuracy of point cloud data matching when using a matching algorithm to match the first point cloud data and the second point cloud data.

[0120] In one embodiment, Figure 5 As shown, a radar calibration verification method is provided, which is described by taking the application of the method to a server as an example, and includes the following steps:

[0121] Step 502: Acquire third point cloud data received by scanning a preset reference plane through a radar to be verified.

[0122] The server obtains third point cloud data received by the radar to be verified scanning a preset reference plane. There may be multiple radars to be verified, and the multiple radars to be verified are calibrated based on the same reference radar. The preset reference plane can be any plane, for example, a flat ground surface or a wall. In one possible implementation, the server obtains third point cloud data received by the multiple radars to be verified simultaneously scanning the preset reference plane.

[0123] Step 504 : convert the third point cloud data received by each radar to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; wherein the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method.

[0124] The radar calibration method described above generates an extrinsic parameter matrix for the radar to be verified. The extrinsic parameter matrix represents the transformation relationship between the coordinate system of the radar to be calibrated and the world coordinate system. Based on the extrinsic parameter matrix, the server can transform the third point cloud data received by all radars to the same coordinate system, generating the fourth point cloud data.

[0125] Step 506 : Determine whether the radar calibration result is accurate based on the flatness of the fourth point cloud data.

[0126] The third point cloud data received by the radar under verification is converted into the same coordinate system to obtain fourth point cloud data corresponding to the third point cloud data received by the radar under verification. The accuracy of the radar calibration result is determined based on the flatness of the fourth point cloud data. The radar calibration result can typically be the extrinsic parameter matrix corresponding to the radar under verification. If the flatness of the fourth point cloud data is greater than a flatness threshold, the radar calibration result is determined to be accurate. If the flatness of the fourth point cloud data is less than the flatness threshold, the radar calibration result is determined to be inaccurate.

[0127] In one possible implementation, the flatness of the fourth point cloud data can be determined based on the range of the coordinates of the fourth point cloud data on the same coordinate axis in the same coordinate system. For example, if the coordinate ranges of the fourth point cloud data on the X-axis and Y-axis are both larger than the coordinate range on the Z-axis, the flatness of the fourth point cloud data is determined based on the coordinate range on the Z-axis. The smaller the coordinate range on the Z-axis, the better the flatness of the fourth point cloud data. Alternatively, a standard reference plane can be established, which is parallel to the plane where the X-axis and Y-axis are located, or can be the plane where the X-axis and Y-axis are located. The distance from each fourth point cloud data to the standard reference plane is calculated, and the flatness of the fourth point cloud data is determined based on the distance range from each fourth point cloud data to the standard reference plane. The smaller the distance range from the fourth point cloud data to the standard reference plane, the better the flatness of the fourth point cloud data.

[0128] In another possible implementation, the normal distribution of the fourth point cloud data corresponding to each radar to be verified is calculated, and the differences between the normal distributions of the fourth point cloud data corresponding to each radar to be verified are compared. Based on the differences between the normal distributions, the flatness of the fourth point cloud data is determined. The smaller the differences between the normal distributions, the better the flatness of the fourth point cloud data.

[0129] The above-mentioned radar calibration verification method obtains the third point cloud data received by the radar to be verified when scanning a preset reference plane; converts the third point cloud data received by each radar to be verified into the same coordinate system according to the external parameter matrix of the radar to be verified, thereby obtaining fourth point cloud data; the external parameter matrix is obtained according to the above-mentioned radar calibration method; and the accuracy of the radar calibration result is determined based on the flatness of the fourth point cloud data. This embodiment converts the third point cloud data received by each radar to be verified into the same coordinate system using the external parameter matrix of the radar to be verified obtained according to the above-mentioned radar calibration method. Based on the flatness of the point cloud data in the same coordinate system, the consistency between the radars to be verified is determined, thereby verifying the accuracy of the radar calibration result, thereby achieving accurate verification of the radar to be verified.

[0130] In one embodiment, Figure 6 As shown, a radar calibration and verification method is provided, comprising the following steps:

[0131] Step 602: Acquire first point cloud data received by a reference radar while moving along a preset route, and second point cloud data received by the radar to be calibrated in the same scenario. In this case, there is at least one position where the visual range of the reference radar includes the visual range of the radar to be calibrated during the movement of the reference radar to the area where the radar to be calibrated is located. Reflectors of different shapes are installed within the visual ranges of different radars to be calibrated. Both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors.

[0132] The first point cloud data is the point cloud data scanned by the reference radar as it moves along a preset route. The second point cloud data is the point cloud data scanned by the radar to be calibrated in the same scenario. The preset route passes through the area where the radar to be calibrated is located and the visual range of the radar to be calibrated, so that the visual range of the reference radar includes the visual range of the radar to be calibrated. At the same time, reflectors of different shapes are installed in the visual ranges of different radars to be calibrated. For example, a circular reflector is installed in the visual range of radar to be calibrated A, a triangular reflector is installed in the visual range of radar to be calibrated B, a square reflector is installed in the visual range of radar to be calibrated C, a pentagonal reflector is installed in the visual range of radar to be calibrated D, and so on.

[0133] Step 604 , matching the first point cloud data and the second point cloud data according to the NDT matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, wherein the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin.

[0134] The extrinsic parameter matrix represents the transformation relationship between the coordinate system of the radar to be calibrated and the reference coordinate system. The reference coordinate system is a coordinate system established with the initial position of the reference radar as its origin. For example, the initial position of the reference radar can be used as the origin. The initial position of the reference radar can be within 3-5 meters of any radar to be calibrated.

[0135] Step 606 : Acquire the third point cloud data received by the radar to be verified scanning the preset reference plane.

[0136] The server obtains third point cloud data received by the radar to be verified scanning the preset reference plane. There may be multiple radars to be verified, and the multiple radars to be verified are calibrated based on the same reference radar. In this embodiment, the radar to be calibrated in steps 602 to 604 can be used as the radar to be verified for accuracy verification of the calibration results.

[0137] Step 608 : Convert the third point cloud data received by each radar to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified to obtain fourth point cloud data; wherein the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method.

[0138] After obtaining the extrinsic parameter matrix of the radar to be verified according to the above radar calibration method, the third point cloud data received by each radar to be verified is converted into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified to obtain the fourth point cloud data.

[0139] Step 610: Determine whether the radar calibration result is accurate based on the flatness of the fourth point cloud data.

[0140] By setting a flatness threshold, if the flatness of the fourth point cloud data is greater than the flatness threshold, it is determined that the radar calibration result is accurate; if the flatness of the fourth point cloud data is less than the flatness threshold, it is determined that the radar calibration result is inaccurate.

[0141] The above radar calibration and verification method can realize the simultaneous calibration of multiple radars to be calibrated through the radar calibration method, and can obtain a calibration result with higher accuracy, that is, the external parameter matrix corresponding to each radar to be calibrated, and then verify the calibration result obtained previously through the radar calibration verification method to achieve accurate verification of the radar calibration result.

[0142] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0143] Based on the same inventive concept, embodiments of the present application also provide a radar calibration device for implementing the aforementioned radar calibration method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more radar calibration device embodiments provided below can be found in the above-described limitations of the radar calibration method and will not be further elaborated here.

[0144] In one embodiment, Figure 7 As shown, a radar calibration device is provided, including: a point cloud acquisition module 702 and a point cloud matching module 704, wherein:

[0145] The point cloud acquisition module 702 is configured to acquire first point cloud data received by a reference radar while it moves along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; wherein, during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position where the visual range of the reference radar includes the visual range of the radar to be calibrated; and wherein reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; and wherein both the first point cloud data and the second point cloud data include point cloud data reflected by the reflectors.

[0146] The point cloud matching module 704 is used to match the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin.

[0147] In one embodiment, the radar calibration device further includes an initialization module, which is configured to:

[0148] Obtaining the initial coordinates of the radar to be calibrated in the reference coordinate system;

[0149] Using the initial coordinates of the radar to be calibrated as initialization matrix parameters of the matching algorithm;

[0150] The point cloud matching module 704 is further configured to:

[0151] The first point cloud data and the second point cloud data are matched using a matching algorithm with the initialization matrix parameters to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

[0152] In one embodiment, the point cloud matching module 704 is further configured to:

[0153] generating map data to be matched based on the first point cloud data;

[0154] Obtaining standard normal distribution parameters of the first point cloud data according to the map data to be matched;

[0155] Obtaining standard normal distribution parameters of the second point cloud data according to the second point cloud data and the initialization matrix parameters;

[0156] An extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system is obtained according to the standard normal distribution parameters of the first point cloud data and the standard normal distribution parameters of the second point cloud data.

[0157] In one embodiment, the point cloud matching module 704 is further configured to:

[0158] Acquire first sub-point cloud data in the first point cloud data, the acquisition time range of which is the same as the acquisition time range of the second point cloud data;

[0159] The first sub-point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

[0160] Based on the same inventive concept, embodiments of the present application also provide a radar calibration verification device for implementing the aforementioned radar calibration verification method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more radar calibration device embodiments provided below can be found in the above-described limitations of the radar calibration verification method and will not be further elaborated here.

[0161] In one embodiment, a radar calibration verification device is provided, comprising:

[0162] A first verification module is used to obtain third point cloud data received by the radar to be verified scanning the preset reference plane;

[0163] A second verification module is configured to convert the third point cloud data received by each of the radars to be verified into a same coordinate system according to an extrinsic parameter matrix of the radar to be verified, to obtain fourth point cloud data; the extrinsic parameter matrix is obtained according to the above-mentioned radar calibration method;

[0164] The third verification module is used to determine whether the radar calibration result is accurate based on the flatness of the fourth point cloud data.

[0165] Each module in the radar calibration device or radar calibration verification device described above may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0166] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. 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, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store point cloud data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a radar calibration method or a radar calibration verification method is implemented.

[0167] Those skilled in the art will understand that Figure 8 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 shown in the figure, or combine certain components, or have a different component arrangement.

[0168] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the radar calibration method in the above embodiment when executing the computer program.

[0169] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the radar calibration method in the above embodiment are implemented.

[0170] In one embodiment, a computer program product is provided, including a computer program, which implements the steps of the radar calibration method in the above embodiment when executed by a processor.

[0171] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the radar calibration verification method in the above embodiment are implemented.

[0172] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the radar calibration verification method in the above embodiment are implemented.

[0173] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the radar calibration verification method in the above embodiment.

[0174] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0175] The technical features of the above embodiments can 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.

[0176] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A radar calibration method, characterized in that: The method comprises: Acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; during the movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position such that the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both include point cloud data reflected by the reflectors; and the first point cloud data includes the second point cloud data received by all radars to be calibrated; The first point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, where the reference coordinate system is a coordinate system established with an initial position of the reference radar as an origin.

2. The method according to claim 1, characterized in that After establishing a reference coordinate system with the initial position of the reference radar as the origin, the method further includes: Obtaining the initial coordinates of the radar to be calibrated in the reference coordinate system; Using the initial coordinates of the radar to be calibrated as initialization matrix parameters of the matching algorithm; The matching of the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system includes: The first point cloud data and the second point cloud data are matched using a matching algorithm with the initialization matrix parameters to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

3. The method according to claim 2, characterized in that The matching algorithm with the initialization matrix parameters is used to match the first point cloud data and the second point cloud data to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, including: generating map data to be matched based on the first point cloud data; Obtaining standard normal distribution parameters of the first point cloud data according to the map data to be matched; Obtaining standard normal distribution parameters of the second point cloud data according to the second point cloud data and the initialization matrix parameters; An extrinsic parameter matrix of the radar to be calibrated in the reference coordinate system is obtained according to the standard normal distribution parameters of the first point cloud data and the standard normal distribution parameters of the second point cloud data.

4. The method according to claim 1, wherein The matching of the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system includes: Acquire first sub-point cloud data in the first point cloud data, the acquisition time range of which is the same as the acquisition time range of the second point cloud data; The first sub-point cloud data and the second point cloud data are matched according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system.

5. The method according to claim 1, wherein The reflective plates include circular reflective plates, triangular reflective plates and polygonal reflective plates.

6. A radar calibration verification method, characterized in that: The method comprises: Acquire third point cloud data received by the radar scanning preset reference plane to be verified; Converting the third point cloud data received by each of the radars to be verified into the same coordinate system according to the extrinsic parameter matrix of the radar to be verified, thereby obtaining fourth point cloud data; wherein the extrinsic parameter matrix is obtained by the radar calibration method according to any one of claims 1 to 5; Whether the radar calibration result is accurate is determined based on the flatness of the fourth point cloud data.

7. A radar calibration device, characterized in that: The device comprises: A point cloud acquisition module is configured to acquire first point cloud data received by a reference radar during movement along a preset route, and second point cloud data received by a radar to be calibrated in the same scenario; wherein, during movement of the reference radar to the area where the radar to be calibrated is located, there is at least one position where the visual range of the reference radar includes the visual range of the radar to be calibrated; reflectors of different shapes are installed within the visual ranges of different radars to be calibrated; the first point cloud data and the second point cloud data both include point cloud data reflected by the reflectors; and the first point cloud data includes the second point cloud data received by all radars to be calibrated; a point cloud matching module, configured to match the first point cloud data and the second point cloud data according to a matching algorithm to obtain an extrinsic parameter matrix of the radar to be calibrated in a reference coordinate system, wherein the reference coordinate system is a coordinate system established with the initial position of the reference radar as the origin.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images