LiDAR intrinsic parameter calibration method, intrinsic and extrinsic parameter calibration method and LiDAR

Abstract

This application discloses an intrinsic parameter calibration method for a lidar, an intrinsic and extrinsic parameter calibration method, and a lidar. The method includes: acquiring multiple frames of point clouds obtained by the lidar scanning a first calibration board and a second calibration board in different attitudes; determining target yaw intrinsic parameters based on the multiple frames of point clouds, wherein the target yaw intrinsic parameters are yaw intrinsic parameter values ​​that enable each frame of point cloud to obtain a consistent inter-board distance value, the yaw intrinsic parameter value is the increment of the yaw angle corresponding to the data point in the point cloud, and the inter-board distance value is the distance from the first calibration board to the second calibration board. This application embodiment, by calibrating the target yaw intrinsic parameters, corrects the point cloud distortion manifested in the yaw angle due to installation errors, which is beneficial for the lidar to obtain reliable and accurate point clouds.

Description

Technical Field

[0001] This application relates to the field of parameter calibration technology, and in particular to a method for calibrating the intrinsic parameters of a lidar, a method for calibrating both intrinsic and extrinsic parameters, and a lidar. Background Technology

[0002] LiDAR (LiDAR) systems contain scanning devices, and the scanning angle of these devices affects the LiDAR's field of view. To expand the field of view, LiDAR systems often have a larger scanning space to allow the scanning devices to increase their scanning angle; however, this approach tends to increase the size of the LiDAR system. Related technologies use wide-angle lenses to magnify the field of view, achieving a larger field of view without increasing the scanning angle. However, installation errors between the scanning devices and the wide-angle lenses can cause distortion in the point cloud data acquired by the LiDAR. Summary of the Invention

[0003] One objective of this application is to provide a method for calibrating the intrinsic and extrinsic parameters of a lidar, as well as a lidar itself, to improve the situation where point cloud distortion occurs due to installation errors between the scanning device and the wide-angle lens in related technologies.

[0004] In a first aspect, embodiments of this application provide an intrinsic parameter calibration method for a lidar, comprising: acquiring multiple frames of point clouds obtained by the lidar scanning a first calibration board and a second calibration board in different attitudes; determining target yaw intrinsic parameters based on the multiple frames of point clouds, wherein the target yaw intrinsic parameters are yaw intrinsic parameter values ​​that enable each frame of point clouds to obtain a consistent inter-board distance value, the yaw intrinsic parameter value is the increment of the yaw angle corresponding to the data point in the point cloud, and the inter-board distance value is the distance from the first calibration board to the second calibration board.

[0005] This application embodiment corrects the point cloud distortion in the yaw angle caused by installation errors by calibrating the target yaw intrinsic parameters, which is beneficial for the lidar to obtain reliable and accurate point clouds.

[0006] Optionally, determining the target yaw intrinsic parameters based on multiple frames of the point cloud includes: configuring multiple yaw intrinsic parameter values ​​for the point cloud in the same frame; updating the point cloud in the same frame using the multiple yaw intrinsic parameter values ​​to obtain multiple inter-board distance values ​​of the point cloud in the same frame; obtaining the correlation between the yaw intrinsic parameter values ​​and the inter-board distance values ​​corresponding to the point cloud in the same frame based on the multiple yaw intrinsic parameter values ​​and the corresponding inter-board distance values; and determining the target yaw intrinsic parameters according to the correlation between the point clouds in each frame.

[0007] Optionally, obtaining the correlation between the yaw intrinsic parameter values ​​and the inter-plate distance values ​​corresponding to the same frame point cloud based on multiple yaw intrinsic parameter values ​​and the corresponding inter-plate distance values ​​includes:

[0008] Linear fitting is performed on multiple yaw intrinsic parameter values ​​of the same frame point cloud and the corresponding inter-board distance values ​​to obtain the first fitted straight line corresponding to the same frame point cloud.

[0009] Optionally, determining the target yaw intrinsic parameters based on the correlation between the multiple frame point clouds includes: determining the yaw intrinsic parameter value corresponding to the intersection point of the first fitted straight line corresponding to each frame of the point cloud as the target yaw intrinsic parameter.

[0010] Optionally, the method further includes: determining target pitch intrinsic parameters based on multiple frames of the point cloud, wherein the target pitch intrinsic parameters are pitch intrinsic parameter values ​​that enable a first absolute difference between the ground height of the first target point cloud and the ground height of the third target point cloud to be equal to a second absolute difference between the ground height of the second target point cloud and the ground height of the third target point cloud, the first field of view corresponding to the first target point cloud and the second field of view corresponding to the second target point cloud to be symmetrical about the third field of view corresponding to the third target point cloud, and the pitch intrinsic parameter value is the increment of the pitch angle corresponding to the data point in the point cloud.

[0011] Optionally, determining the target pitch intrinsic parameters based on multiple frames of the point cloud includes: configuring multiple pitch intrinsic parameter values ​​for the same frame point cloud; updating the same frame point cloud with the multiple pitch intrinsic parameter values ​​respectively to obtain multiple ground heights of the same frame point cloud; obtaining the correlation between the pitch intrinsic parameter values ​​corresponding to the same frame point cloud and the ground height based on the multiple pitch intrinsic parameter values ​​of the same frame point cloud and the corresponding ground heights; and determining the target pitch intrinsic parameters according to the correlation between the point cloud in each frame.

[0012] Optionally, obtaining the correlation between the pitch intrinsic values ​​of the same frame point cloud and the ground height based on multiple pitch intrinsic values ​​of the same frame point cloud and the corresponding ground height includes: performing linear fitting on multiple pitch intrinsic values ​​of the same frame point cloud and the corresponding ground height to obtain a second fitted straight line corresponding to the same frame point cloud.

[0013] Optionally, each frame of the point cloud corresponds to a first pitch line, a second pitch line, and a third pitch line, all of which are second fitted lines. The step of determining the target pitch intrinsic parameter based on the correlation between the point clouds in each frame includes: obtaining a reference pitch intrinsic parameter; determining a first ground height based on the first pitch line and the reference pitch intrinsic parameter; determining a second ground height based on the second pitch line and the reference pitch intrinsic parameter; determining a third ground height based on the third pitch line and the reference pitch intrinsic parameter; calculating a first absolute value of the difference between the first ground height and the second ground height; calculating a second absolute value of the difference between the second ground height and the third ground height; and determining the reference pitch intrinsic parameter as the target pitch intrinsic parameter in response to the first absolute value being equal to the second absolute value.

[0014] Optionally, the method further includes: determining a target roll intrinsic parameter based on multiple frames of the point cloud, wherein the target roll intrinsic parameter is a roll intrinsic parameter value that enables the point cloud in each frame to obtain a consistent ground height, and the roll intrinsic parameter value is the increment of the roll angle corresponding to the data point in the point cloud.

[0015] Optionally, determining the target roll intrinsic parameter based on multiple frames of the point cloud includes: configuring multiple roll intrinsic parameter values ​​for the point cloud in the same frame; updating the point cloud in the same frame using the multiple roll intrinsic parameter values ​​respectively to obtain multiple ground heights of the point cloud in the same frame; obtaining the correlation between the roll intrinsic parameter values ​​and the ground heights corresponding to the point cloud in the same frame based on the multiple roll intrinsic parameter values ​​and the corresponding ground heights; and determining the target roll intrinsic parameter according to the correlation between the point cloud in each frame.

[0016] Optionally, obtaining the correlation between the roll intrinsic values ​​of the same frame point cloud and the corresponding ground height based on multiple roll intrinsic values ​​of the same frame point cloud and the corresponding ground height includes: performing linear fitting on multiple roll intrinsic values ​​of the same frame point cloud and the corresponding ground height to obtain a third fitted straight line corresponding to the same frame point cloud.

[0017] Optionally, determining the target roll intrinsic parameter based on the correlation relationship between the point clouds in each frame includes: determining the roll intrinsic parameter value corresponding to the intersection point of the third fitted line corresponding to each point cloud as the target roll intrinsic parameter.

[0018] Optionally, the field of view range of the lidar is [m, n]. The step of acquiring multiple frames of point cloud obtained by the lidar scanning the first and second calibration boards in different orientations includes: acquiring the point cloud obtained by the lidar scanning the first and second calibration boards in a first orientation, where the first orientation is the orientation of the lidar when the laser line at the m-th field of view hits the first calibration board; acquiring the point cloud obtained by the lidar scanning the first and second calibration boards in a second orientation, where the second orientation is the orientation of the lidar when the laser line at the n-th field of view hits the second calibration board; and acquiring the point cloud obtained by the lidar scanning the first and second calibration boards in a third orientation, where the third orientation is the orientation of the lidar when the laser line at the n-th field of view hits the second calibration board. The posture of the laser line in the field of view when it hits the middle of the first calibration plate and the second calibration plate.

[0019] In a second aspect, embodiments of this application provide a method for calibrating the intrinsic and extrinsic parameters of a lidar, comprising: determining candidate extrinsic parameter data of the lidar, the candidate extrinsic parameter data including candidate yaw extrinsic parameters, candidate pitch extrinsic parameters and candidate roll extrinsic parameters; and obtaining target intrinsic parameter data based on the above-described intrinsic parameter calibration method, the target intrinsic parameter data including target yaw intrinsic parameters, target pitch intrinsic parameters and target roll intrinsic parameters.

[0020] Optionally, the candidate extrinsic parameter data includes candidate yaw extrinsic parameters. Determining the candidate extrinsic parameter data of the lidar includes: acquiring ground truth data, which represents the ground truth of the calibration site, wherein the first calibration plate and the second calibration plate are spaced apart on the calibration site; determining a first normal vector of the ground truth based on the ground truth data; determining a second normal vector of the actual ground based on the point cloud, wherein the actual ground is the ground of the calibration site detected by the lidar; and determining a first angle between the first normal vector and the second normal vector, wherein the first angle is the candidate yaw extrinsic parameter.

[0021] Optionally, the candidate extrinsic parameter data includes candidate pitch extrinsic parameters. Determining the candidate extrinsic parameter data of the lidar includes: acquiring ground truth panel data, the ground truth panel data being used to represent the ground truth panel corresponding to the first calibration panel and the second calibration panel; determining a third normal vector of the ground truth panel based on the ground truth panel data; determining a fourth normal vector of the actual panel based on the point cloud, the actual panel being the panel obtained by the lidar detecting the first calibration panel or the second calibration panel; and determining a second included angle between the third normal vector and the fourth normal vector, the second included angle being the candidate pitch extrinsic parameter.

[0022] Optionally, the candidate extrinsic parameter data includes candidate roll extrinsic parameters. Determining the candidate extrinsic parameter data of the lidar includes: acquiring ground truth data and ground truth panel data, wherein the ground truth data represents the ground truth of the calibration site, and the ground truth panel data represents the ground truth panel corresponding to the first calibration panel and the second calibration panel; determining a first normal vector of the ground truth based on the ground truth data; determining a third normal vector of the ground truth based on the ground truth panel data; performing a cross product of the first normal vector and the third normal vector to obtain a fifth normal vector; determining a second normal vector of the actual ground and a fourth normal vector of the first calibration panel or the second calibration panel based on the point cloud, wherein the actual ground is the ground of the calibration site detected by the lidar; performing a cross product of the second normal vector and the fourth normal vector to obtain a sixth normal vector; and determining a third angle between the fifth normal vector and the sixth normal vector, wherein the third angle is the candidate roll extrinsic parameter.

[0023] Optionally, the method further includes: updating the candidate extrinsic data based on the target intrinsic data to obtain target extrinsic data, wherein the target extrinsic data includes target yaw extrinsic data, target pitch extrinsic data, and target roll extrinsic data.

[0024] Optionally, updating the candidate extrinsic data based on the target intrinsic data to obtain the target extrinsic data includes: subtracting the candidate yaw extrinsic data from the target yaw intrinsic data by a preset multiple to obtain the target yaw extrinsic data; subtracting the candidate pitch extrinsic data from the target pitch intrinsic data by a preset multiple to obtain the target pitch extrinsic data; and subtracting the candidate roll extrinsic data from the target roll intrinsic data by a preset multiple to obtain the target roll extrinsic data.

[0025] In a third aspect, embodiments of this application provide a lidar, including a memory and a processor. The memory is connected to the processor, and the processor is configured to execute one or more computer programs stored in the memory. When the processor executes the one or more computer programs, it causes the lidar to implement the aforementioned lidar intrinsic parameter calibration method or the aforementioned lidar intrinsic and extrinsic parameter calibration method.

[0026] In a fourth aspect, embodiments of this application provide a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, cause the processor to perform the aforementioned intrinsic parameter calibration method for a lidar or the aforementioned intrinsic and extrinsic parameter calibration method for a lidar. Attached Figure Description

[0027] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 A schematic diagram of a lidar system architecture provided for related technologies;

[0029] Figure 2 A schematic diagram of a calibration location provided for an embodiment of this application;

[0030] Figure 3 This is a flowchart illustrating a method for calibrating the extrinsic parameters of a lidar, as provided in an embodiment of this application. Figure 3 The method shown is used to calibrate candidate yaw extrinsic parameters;

[0031] Figure 4This is a flowchart illustrating a method for calibrating the extrinsic parameters of a lidar, as provided in an embodiment of this application. Figure 4 The method shown is used to calibrate candidate pitch extrinsic parameters;

[0032] Figure 5 A flowchart illustrating a method for calibrating the extrinsic parameters of a lidar, as provided in another embodiment of this application, is shown below. Figure 5 The method shown is used to calibrate candidate roll extrinsic parameters;

[0033] Figure 6 A flowchart illustrating an intrinsic parameter calibration method for a lidar provided in this application embodiment is shown below. Figure 6 The method shown is used to calibrate the target yaw intrinsic parameters;

[0034] Figure 7 A schematic diagram of a scenario in which a lidar scans a first calibration board and a second calibration board in a first posture, provided for an embodiment of this application;

[0035] Figure 8 A schematic diagram of a scenario in which a lidar scans a first calibration board and a second calibration board in a second posture, provided as an embodiment of this application;

[0036] Figure 9 A schematic diagram illustrating a scenario in which a lidar scans a first calibration board and a second calibration board in a third posture, as provided in an embodiment of this application;

[0037] Figure 10 The left edge point cloud map is updated using a first yaw intrinsic parameter value and a second yaw intrinsic parameter value, as provided in the embodiments of this application.

[0038] Figure 11 The embodiments of this application provide for updating the right edge point cloud using a first yaw intrinsic parameter value and a second yaw intrinsic parameter value;

[0039] Figure 12 The embodiments of this application provide for updating the left edge point cloud using a first roll intrinsic value and a second roll intrinsic value;

[0040] Figure 13 The embodiments of this application provide for updating the right edge point cloud using a first roll intrinsic value and a second roll intrinsic value;

[0041] Figure 14 A schematic diagram of the structure of an internal parameter calibration device for a lidar provided in an embodiment of this application;

[0042] Figure 15 A schematic diagram of the internal and external parameter calibration device for a lidar provided in this application embodiment;

[0043] Figure 16This is a schematic diagram of the structure of a lidar provided in an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0045] It should be noted that, unless there is a conflict, the various features in the embodiments of this application can be combined with each other, all of which are within the protection scope of this application. Furthermore, although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than the module division in the device or the order in the flowchart. Moreover, the terms "first," "second," and "third" used in this application do not limit the data or execution order, but only distinguish identical or similar items with essentially the same function and effect.

[0046] I. Overview of Related Technologies

[0047] LiDAR scanning devices are divided into one-dimensional scanning devices and two-dimensional scanning devices. For example, a one-dimensional scanning device may include a rotating mirror, and a two-dimensional scanning device may include a MEMS galvanometer. Both one-dimensional and two-dimensional scanning devices are linear scanning devices, and the laser line is magnified according to the optical magnification after passing through the expanding lens.

[0048] Please see Figure 1 The galvanometer 11 and the expanding lens 12 are mounted opposite each other. The laser line 13 is incident on the reflecting mirror 14 and, after reflection by the reflecting mirror 14, is incident on the galvanometer 11. The galvanometer 11 reflects the laser line 13 to the expanding lens 12, which magnifies the exit angle of the laser line 13 according to a preset optical magnification. Ideally, the center of the galvanometer 11 and the center of the expanding lens 12 are on the same horizontal line, so that the laser line reflected by the galvanometer 11 is orthogonal to the exit surface of the expanding lens 12. However, in reality, there is an installation error between the galvanometer 11 and the expanding lens 12. This installation error will cause a large difference between the light exit angle at large angles and the ideal angle, resulting in distortion of the point cloud collected by the lidar.

[0049] II. Analysis of the Principles Behind Distortion in Point Clouds Acquired by LiDAR

[0050] The position of each data point in the point cloud of the lidar is related to the yaw angle and the pitch angle, respectively, as shown in Formula 1, as follows:

[0051]

[0052] Formula 1

[0053]

[0054] The lidar is equipped with a three-dimensional coordinate system XYZ, the position of the data point is (x,y,z), and dist is the distance from the data point to the origin of the three-dimensional coordinate system XYZ.

[0055] There is an installation error between the galvanometer 11 and the expansion lens 12, which causes errors in the yaw angle and / or pitch angle. As can be seen from Formula 1, this makes the position of each data point calculated by the lidar inaccurate. For example, when describing the shape of a cuboid object based on the point cloud with errors, the shape of the object is deformed and is no longer described as a cuboid. Therefore, the object exhibits point cloud distortion.

[0056] It is understandable that the expanding lens amplifies the angle of the laser line reflected by the galvanometer before it is emitted. When there is an installation error between the galvanometer 11 and the expanding lens 12, the larger the field of view, the more severe the distortion of the data point corresponding to that field of view; conversely, the smaller the field of view, the less severe the distortion of the data point corresponding to that field of view. For example, if the scanning angle range of the galvanometer is [-30°, 30°], after being magnified twice by the expanding lens, the field of view range of the lidar is [-60°, 60°]. The distortion of the data point corresponding to the -60° field of view is greater than that of the data point corresponding to the -59° field of view, and similarly, the distortion of the data point corresponding to the 60° field of view is greater than that of the data point corresponding to the 59° field of view. Therefore, the maximum distortion occurs at the data points corresponding to the edge of the field of view range, such as the data point corresponding to the -60° or 60° field of view.

[0057] III. Construction of the calibration site provided in the embodiments of this application

[0058] Please see Figure 2In this embodiment, a calibration site is constructed, in which a first calibration plate 21, a second calibration plate 22, and a turntable 23 are set. A lidar 24 is placed on the turntable 23, which rotates the lidar 24, changing its orientation so that the lidar 24 emits laser lines to scan the first calibration plate 21 and the second calibration plate 22. A portion of the laser lines in the field of view hits the first calibration plate 21 and / or the second calibration plate 22 and is reflected back to the lidar 24. The remaining portion of the laser lines in the field of view do not hit the first calibration plate 21 and / or the second calibration plate 22 but hit other objects or the ground, and are reflected back to the lidar 24.

[0059] It is understood that, in the embodiments of this application, the rotation of the turntable 23 can be controlled, the turntable 23 can stop at multiple angular positions, and the lidar can be controlled to scan a point cloud at the corresponding position. In this way, the laser lines of different field of view of the lidar 24 can hit the first calibration plate 21 and / or the second calibration plate 22, and be returned to the lidar by the first calibration plate 21 and / or the second calibration plate 22. Thus, the internal and external parameters of different field of view within the field of view of the lidar can be calibrated.

[0060] IV. Overview of the Implementation of the Embodiments of this Application

[0061] The overall implementation process of this application's embodiment is as follows:

[0062] 4.1) Control the turntable to drive the lidar to acquire multiple frames of point cloud data in different postures;

[0063] 4.2) Determine the candidate extrinsic parameters of the lidar. The candidate extrinsic parameters include candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters. The candidate extrinsic parameters are used to represent the current attitude of the lidar.

[0064] 4.3) Obtain target intrinsic parameter data, including target yaw intrinsic parameters, target pitch intrinsic parameters, and target roll intrinsic parameters;

[0065] 4.4) Update the candidate extrinsic data with the target intrinsic data to obtain the target extrinsic data, which includes the target yaw extrinsic data, the target pitch extrinsic data and the target roll extrinsic data.

[0066] 4.5) Verify whether the target intrinsic parameter data meets the requirements.

[0067] It is understood that the content of 4.1) has already been discussed in point ③, and will not be repeated here. The following embodiments of this application will be discussed in relation to 4.2) to 4.5).

[0068] V. Determining Candidate External Parameter Data for LiDAR

[0069] Candidate extrinsic parameter data include candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters. This application's embodiments describe the determination of candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters as follows:

[0070] 5.1) Determine candidate yaw external parameters.

[0071] Please see Figure 3 In this embodiment of the application, candidate yaw extrinsic parameters are determined through steps S31 to S34, as detailed below:

[0072] Step S31: Obtain true ground data.

[0073] The true ground data is used to represent the true ground of the calibration site, and the first calibration plate and the second calibration plate are set at intervals on the calibration site. In this embodiment, a high-precision lidar is used to pre-detect the ground of the calibration site to obtain the true ground data.

[0074] Step S32: Determine the first normal vector of the true ground based on the true ground data.

[0075] The first normal vector is the normal vector perpendicular to the true ground. The yaw angle is the angle by which the lidar rotates around the axis perpendicular to the actual ground. The yaw angle of the lidar in the current attitude can be represented by the angle between the normal vector of the actual ground obtained by the lidar and the normal vector of the true ground.

[0076] This application's embodiments determine the plane equation of the true ground based on true ground data, and determine the first normal vector of the true ground based on the plane equation of the true ground. .

[0077] Step S33: Determine the second normal vector of the actual ground based on the point cloud.

[0078] The actual ground is the ground at the calibration location detected by the lidar to be calibrated, and the second normal vector is a normal vector perpendicular to the actual ground. In this embodiment, actual ground data corresponding to the actual ground is extracted from the point cloud, a plane equation for the actual ground is determined based on the actual ground data, and a second normal vector for the actual ground is determined based on the plane equation of the actual ground. .

[0079] Step S34: Determine the first angle between the first normal vector and the second normal vector.

[0080] In this embodiment of the application, the first included angle is obtained by performing a dot product of the first normal vector and the second normal vector according to Formula 2, as shown below:

[0081] Formula 2

[0082] Among them, the first included angle The candidate yaw external parameter.

[0083] It is understandable that when a lidar is configured with a three-dimensional coordinate system XYZ, with the X-axis perpendicular to the lidar's output surface, the Y-axis perpendicular to the X-axis in the horizontal plane, and the Z-axis perpendicular to the Y-axis in the vertical plane, the yaw angle is the angle by which the lidar rotates around the Z-axis.

[0084] 5.2) Determine candidate pitch extrinsic parameters.

[0085] Please see Figure 4 In this embodiment of the application, candidate yaw extrinsic parameters are determined through steps S41 to S44, as detailed below:

[0086] Step S41: Obtain the truth panel data.

[0087] The truth panel data is used to represent the truth panel corresponding to the first calibration board and the second calibration board. In this embodiment, a high-precision lidar is used to pre-detect the first and second calibration boards to obtain the truth panel data.

[0088] Step S42: Determine the third normal vector of the truth panel based on the truth panel data.

[0089] The third normal vector is the normal vector perpendicular to the true value plate surface. The pitch angle is the angle by which the lidar rotates around the axis perpendicular to the first or second calibration plate. The pitch angle of the lidar in the current attitude can be represented by the angle between the normal vector of the actual plate surface obtained by the lidar and the normal vector of the true value plate surface.

[0090] This application's embodiments determine the plane equation about the truth plate surface based on the truth plate surface data, and determine the third normal vector of the truth plate surface based on the plane equation of the truth plate surface. .

[0091] Step S43: Determine the fourth normal vector of the actual board surface based on the point cloud.

[0092] The actual board surface is the surface obtained by the lidar to be calibrated from detecting the first or second calibration board, and the fourth normal vector is a normal vector perpendicular to the actual board surface. In this embodiment, actual board surface data corresponding to the actual board surface is extracted from the point cloud, a plane equation about the actual board surface is determined based on the actual board surface data, and the fourth normal vector of the actual board surface is determined based on the plane equation of the actual board surface. .

[0093] Step S44: Determine the second angle between the third normal vector and the fourth normal vector.

[0094] In this embodiment of the application, the third normal vector and the fourth normal vector are multiplied by formula 3 to obtain the second included angle, as shown below:

[0095] Formula 2

[0096] Among them, the second included angle The pitch parameter is a candidate pitch extrinsic parameter.

[0097] It is understandable that when a lidar is configured with a three-dimensional coordinate system XYZ, with the X-axis perpendicular to the lidar's output surface, the Y-axis perpendicular to the X-axis in the horizontal plane, and the Z-axis perpendicular to the Y-axis in the vertical plane, the pitch angle is the angle by which the lidar rotates around the Y-axis.

[0098] 5.3) Determine candidate rolling extrinsic parameters.

[0099] Please see Figure 5 In this embodiment of the application, candidate roll extrinsic parameters are determined through steps S51 to S57, as detailed below:

[0100] Step S51: Obtain ground truth data and ground truth panel data.

[0101] Step S52: Determine the first normal vector of the true ground based on the true data.

[0102] Step S53: Determine the third normal vector of the truth panel based on the truth panel data.

[0103] Step S54: Perform a cross product of the first normal vector and the third normal vector to obtain the fifth normal vector.

[0104] Step S55: Determine the second normal vector of the actual ground and the fourth normal vector of the first or second calibration plate based on the point cloud. The actual ground is the ground of the calibration site detected by the lidar.

[0105] Step S56: Perform a cross product of the second normal vector and the fourth normal vector to obtain the sixth normal vector.

[0106] Step S57: Determine the third angle between the fifth normal vector and the sixth normal vector. The third angle is a candidate roll extrinsic parameter.

[0107] In steps S51 to S53, the methods for obtaining the true ground data and the true plate data, as well as the methods for determining the first normal vector and the third normal vector, have been described above and will not be repeated here.

[0108] In step S54, the fifth normal vector is perpendicular to the plane formed by the first and second normal vectors. In this embodiment, the fifth normal vector is obtained according to Formula 3, as shown below:

[0109] Formula 3

[0110] in, This is the fifth normal vector.

[0111] In step S55, the methods for obtaining the second and fourth normal vectors have been discussed above and will not be repeated here.

[0112] In step S56, the sixth normal vector is perpendicular to the plane formed by the second and fourth normal vectors. In this embodiment, the sixth normal vector is obtained according to Formula 4, as shown below:

[0113] Formula 4

[0114] in, It is the sixth normal vector.

[0115] In step S57, the candidate roll extrinsic parameters can be represented by the angle between the fifth normal vector and the sixth normal vector. In this embodiment, the candidate roll extrinsic parameters are obtained according to Formula 5, as shown below:

[0116] Formula 5

[0117] in, The third included angle, the third included angle For candidate rolling external parameters.

[0118] It is understandable that when a lidar is configured with a three-dimensional coordinate system XYZ, with the X-axis perpendicular to the lidar's output surface, the Y-axis perpendicular to the X-axis in the horizontal plane, and the Z-axis perpendicular to the Y-axis in the vertical plane, the roll angle is the angle by which the lidar rotates around the X-axis.

[0119] Thus far, the embodiments of this application have described the process of determining candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters.

[0120] VI. Obtaining target internal parameter data

[0121] The target intrinsic data includes target yaw intrinsic, target pitch intrinsic, and target roll intrinsic. This application's embodiments describe the determination of the target yaw intrinsic, target pitch intrinsic, and target roll intrinsic, as follows:

[0122] 6.1) Determine the target yaw internal parameters

[0123] Please see Figure 6 In this embodiment of the application, the target yaw intrinsic parameters are determined through steps S61 to S62, as detailed below:

[0124] Step S61: Obtain multiple frames of point cloud data obtained by scanning the first calibration board and the second calibration board with the lidar in different postures.

[0125] In some embodiments, the point cloud is obtained by scanning a first calibration board and a second calibration board with a lidar in any posture. The lidar outputs a laser line to scan the calibration board within the field of view of that posture, where the field of view is the range of angles that the lidar can scan. The field of view of the lidar is [m, n]. For example, m = -30°, n = 30°, and the field of view is [-30°, 30°], where the field of view is an angle in [-30°, 30°]. Another example is m = 0°, n = 60°, and the field of view is [0°, 60°].

[0126] For example, the point cloud is obtained by scanning the first and second calibration boards with a lidar in a reference attitude. The reference attitude is the attitude when the laser line at the k-th field of view of the lidar can hit the first or second calibration board, where k is not equal to m and not equal to n. For example, m = -30°, n = 30°, the field of view range is [-30°, 30°], and k is not equal to -30° and not equal to 30°, such as k = -29° or 0°.

[0127] As mentioned earlier, when there is an installation error between the galvanometer and the expanding lens, the larger the field of view, the more severe the distortion of the data points corresponding to that field of view; conversely, the smaller the field of view, the less severe the distortion of the data points corresponding to that field of view. To obtain more accurate target yaw, pitch, and roll intrinsic parameters, in some embodiments, when the laser line at the edge of the field of view hits the first or second side plate, the current point cloud is acquired. The edge of the field of view is either the minimum or maximum field of view within the field of view range.

[0128] Specifically, in some other embodiments, the point cloud is obtained by scanning the first calibration board and the second calibration board with a lidar in a specified posture. In this embodiment, multiple frames of point cloud are obtained through steps 611 to 613, as detailed below:

[0129] Step S611: Obtain the point cloud obtained by the lidar scanning the first calibration board and the second calibration board in the first posture. The first posture is the posture when the laser line of the lidar at the mth field of view hits the first calibration board.

[0130] The turntable can change the orientation of the lidar, allowing laser lines from different field-of-view angles to hit the first calibration plate. To obtain a point cloud with maximum distortion, the lidar scans both the first and second calibration plates in its first orientation to obtain a point cloud. Specifically, when the lidar's laser line at the m-th field-of-view angle hits the first calibration plate, the lidar scans both plates to obtain a point cloud (which can be named the left edge point cloud). The left edge point cloud is the point cloud collected when the lidar's laser line at its minimum field-of-view angle hits the first calibration plate. The m-th field-of-view angle is the minimum field-of-view angle within the lidar's field-of-view range; for example, if m is -30°, the m-th field-of-view angle is the field-of-view angle at -30°.

[0131] For example, please refer to Figure 7 The field of view range is [-30°, 30°]. When the laser line 71 of the -30° field of view of the lidar hits the first calibration plate 21, the lidar scans the first calibration plate 21 and the second calibration plate 22 to obtain the left edge point cloud.

[0132] Step S612: Obtain the point cloud obtained by the lidar scanning the first calibration board and the second calibration board in the second posture. The second posture is the posture when the laser line of the nth field of view of the lidar hits the second calibration board.

[0133] To obtain another point cloud with the greatest distortion, the lidar scans the first and second calibration boards in a second posture to obtain the point cloud. Specifically, when the lidar's laser line at its nth field of view hits the second calibration board, the lidar scans both the first and second calibration boards, thus obtaining the point cloud (which can be named the right edge point cloud). The right edge point cloud is the point cloud collected by the lidar when its laser line at its maximum field of view hits the first calibration board. The nth field of view is the maximum field of view within the lidar's field of view range; for example, if n is 30°, the nth field of view is the 30° field of view.

[0134] For example, please refer to Figure 8 The field of view range is [-30°, 30°]. When the laser line 72 of the 30° field of view of the lidar hits the second calibration plate 22, the lidar scans the first calibration plate 21 and the second calibration plate 22 to obtain the right edge point cloud.

[0135] Step S613: Obtain the point cloud obtained by the lidar scanning the first and second calibration boards in the third pose. The third pose is the lidar's... The posture of the laser line in the field of view when it hits the middle of the first calibration plate and the second calibration plate.

[0136] To obtain another point cloud with minimal distortion, the lidar scans the point cloud obtained from the first and second calibration boards in a third pose, i.e.: when the lidar's first... When the laser line in the field of view hits the middle of the first calibration plate and the second calibration plate, the lidar scans the first calibration plate and the second calibration plate to obtain a point cloud (which can be named the intermediate point cloud). The intermediate point cloud is the point cloud collected by the lidar when the laser line in the middle field of view hits the middle of the first calibration plate and the second calibration plate.

[0137] For example, please refer to Figure 9 The field of view range is [-30°, 30°]. When the laser line 73 of the 0° field of view of the lidar hits the middle between the first calibration plate 21 and the second calibration plate 22, the lidar scans the first calibration plate 21 and the second calibration plate 22 to obtain the middle point cloud.

[0138] Based on the left edge point cloud, right edge point cloud and middle point cloud, the embodiments of this application can obtain accurate and reliable target yaw intrinsic parameters, target pitch intrinsic parameters and target roll intrinsic parameters.

[0139] Step S62: Determine the target yaw intrinsic parameters based on multi-frame point cloud.

[0140] The target yaw intrinsic parameter is a yaw intrinsic parameter value that enables each frame of the point cloud to obtain a consistent inter-plate distance value. The inter-plate distance value is the distance from the first calibration plate to the second calibration plate, and the yaw intrinsic parameter value is the increment of the yaw angle corresponding to the data point in the point cloud. Based on the yaw intrinsic parameter value, the expression of the yaw angle in this embodiment is modified, as shown in Formula 6:

[0141] Formula 6

[0142] in, This is the yaw internal parameter value. As a candidate yaw external parameter, This is the new yaw extrinsic parameter obtained after updating the candidate yaw extrinsic parameters based on the yaw intrinsic parameter value.

[0143] This application embodiment, by modifying the yaw intrinsic parameter value, can affect the position (x, y, z) of each data point in the point cloud, thereby achieving point cloud updates. This application embodiment uses Equation 7 to illustrate the influence of the yaw intrinsic parameter value on the position (x, y, z) of the data points, as shown below:

[0144]

[0145]

[0146] Formula 7

[0147]

[0148]

[0149]

[0150] Among them, position ( , , ) represents the new location of the data point. The magnification angle is the angle after the new yaw external parameters have been processed by the magnification lens. The zoom angle is the angle after the new pitch extrinsic parameters have been processed by the zoom lens. The new pitch extrinsic parameter is obtained after updating the candidate pitch extrinsic parameters based on the pitch intrinsic parameter value. dist is the distance from the data point to the origin of the three-dimensional coordinate system of the lidar.

[0151] Combining formulas 6 and 7, it can be seen that by modifying the yaw intrinsic parameter value... This achieves the purpose of modifying candidate yaw extrinsic parameters. When the candidate yaw extrinsic parameters are modified, the position (x, y, z) of the data point is recalculated, thereby realizing the update of the point cloud.

[0152] As mentioned above, there are installation errors in the galvanometer and the expansion lens, which cause the distance between the plates detected by the lidar based on different yaw angles to be inconsistent. The embodiments of this application determine the target yaw intrinsic parameters based on multiple frames of point cloud. The target yaw intrinsic parameters can correct the point cloud distortion that appears in the yaw angle due to installation errors, so that the lidar can obtain a consistent distance value between the plates based on each frame of point cloud.

[0153] In this embodiment, the target yaw intrinsic parameters are determined through steps S621 to S624, as shown below:

[0154] Step S621: Configure multiple yaw intrinsic parameter values ​​for the point cloud in the same frame.

[0155] A point cloud within the same frame refers to a point cloud in which multiple yaw intrinsic parameter values ​​are applied to the same frame. This point cloud can be a left-edge point cloud or a right-edge point cloud. In this embodiment, multiple yaw intrinsic parameter values ​​are configured for the same frame point cloud. For example, this embodiment configures a first yaw intrinsic parameter value w1 and a second yaw intrinsic parameter value w2 for the left-edge point cloud, where w1 can be 0 and w2 can be t.

[0156] Step S622: Update the point cloud in the same frame using multiple yaw intrinsic parameter values ​​to obtain multiple inter-board distance values ​​of the point cloud in the same frame.

[0157] In this embodiment, each data point in the same frame point cloud is updated using Formula 7 by combining each yaw intrinsic parameter value, thereby completing the update of the same frame point cloud. It can be understood that when the first yaw intrinsic parameter value w1=0, the same frame point cloud remains unchanged.

[0158] Multiple yaw intrinsic values ​​include a first yaw intrinsic value and a second yaw intrinsic value. The point cloud in the same frame can be a left edge point cloud or a right edge point cloud.

[0159] Please refer to the following: Figure 10 and Figure 11 In this embodiment, the left edge point cloud is updated using a first yaw intrinsic parameter value to obtain a first point cloud, and the left edge point cloud is updated using a second yaw intrinsic parameter value to obtain a second point cloud. In this embodiment, the right edge point cloud is updated using the first yaw intrinsic parameter value to obtain a third point cloud, and the right edge point cloud is updated using the second yaw intrinsic parameter value to obtain a fourth point cloud. In this embodiment, a first inter-plate distance value d11 is determined based on the first point cloud, a second inter-plate distance value d21 is determined based on the second point cloud, a third inter-plate distance value d12 is determined based on the third point cloud, and a fourth inter-plate distance value d22 is determined based on the fourth point cloud.

[0160] Step S623: Based on multiple yaw intrinsic parameter values ​​of the same frame point cloud and the corresponding inter-board distance values, obtain the correlation between the yaw intrinsic parameter values ​​and the inter-board distance values ​​of the same frame point cloud.

[0161] This application embodiment performs linear fitting on multiple yaw intrinsic parameter values ​​of the same frame point cloud with the corresponding inter-plate distance values ​​to obtain a first fitted straight line corresponding to the same frame point cloud. Specifically, this application embodiment generates a first fitted straight line L11 based on the first yaw intrinsic parameter value w1 and the first inter-plate distance value d11, and the second yaw intrinsic parameter value w2 and the second inter-plate distance value d21. It also generates a first fitted straight line L12 based on the first yaw intrinsic parameter value w1 and the third inter-plate distance value d12, and the second yaw intrinsic parameter value w2 and the fourth inter-plate distance value d22.

[0162] The first fitted line L11 describes the relationship between the yaw intrinsic parameter value and the inter-plate distance value of the left edge point cloud under different yaw intrinsic parameter values. The first fitted line L12 describes the relationship between the yaw intrinsic parameter value and the inter-plate distance value of the right edge point cloud under different yaw intrinsic parameter values.

[0163] Both the first fitted line L11 and the first fitted line L12 are straight lines with the yaw internal parameter as the independent variable and the inter-plate distance as the dependent variable.

[0164] The first fitted line L11 is represented by a first linear function, which is calculated as follows: the first linear function y=k1x+b1 is constructed in advance, and the first fitted point (w1,d11) and the second fitted point (w2,d21) are substituted into the first linear function y=k1x+b1 to obtain k1 and b1, and then the expression of the first linear function is obtained.

[0165] The first fitted line L12 is represented by a second linear function, which is calculated as follows: the second linear function y=k2x+b2 is constructed in advance. The third fitting point (w1,d12) and the fourth fitting point (w2,d22) are substituted into the second linear function y=k2x+b2 to obtain k2 and b2, and then the expression of the second linear function is obtained.

[0166] Step S624: Determine the target yaw intrinsic parameters based on the correlation between the point clouds in each frame.

[0167] In this embodiment, the yaw intrinsic parameter value corresponding to the intersection point of the first fitted straight line corresponding to each frame point cloud is determined as the target yaw intrinsic parameter. Specifically, in this embodiment, the yaw intrinsic parameter value corresponding to the intersection point of the first fitted straight line L1 and the first fitted straight line L2 is determined as the target yaw intrinsic parameter.

[0168] The target yaw intrinsic parameter ensures that the lidar obtains the same distance result when observing the distance between the first and second calibration plates at different yaw angles. Therefore, the target yaw intrinsic parameter corrects the point cloud distortion caused by installation errors in the yaw angle direction. Furthermore, the embodiments of this application can obtain accurate and reliable target yaw intrinsic parameters by relying on only a few frames of point cloud (left edge point cloud and right edge point cloud), which is efficient and accurate.

[0169] 6.2) Determine the target pitch parameters

[0170] The embodiments of this application determine the target pitch intrinsic parameters through the following steps, as shown below: Based on multiple frame point clouds, determine the target pitch intrinsic parameters.

[0171] The target pitch intrinsic parameter is a pitch intrinsic parameter value that makes the first absolute difference between the ground height of the first target point cloud and the ground height of the third target point cloud equal to the second absolute difference between the ground height of the second target point cloud and the ground height of the third target point cloud. The first field of view corresponding to the first target point cloud and the second field of view corresponding to the second target point cloud are symmetrical about the third field of view corresponding to the third target point cloud.

[0172] For example, the first target point cloud is the point cloud collected by the lidar when scanning the first calibration board and the second calibration board with a first field of view of [-50°, 10°], the second target point cloud is the point cloud collected by the lidar when scanning the first calibration board and the second calibration board with a second field of view of [-10°, 50°], and the third target point cloud is the point cloud collected by the lidar when scanning the first calibration board and the second calibration board with a third field of view of [-30°, 30°]. The first field of view and the second field of view are symmetrical about the third field of view.

[0173] In some embodiments, the first target point cloud is the left edge point cloud, the second target point cloud is the right edge point cloud, and the third target point cloud is the middle point cloud.

[0174] The pitch intrinsic parameter value is the increment of the pitch angle corresponding to the data point in the point cloud. Based on the pitch intrinsic parameter value, this application embodiment modifies the expression of the pitch angle, wherein the expression of the pitch angle is: ,in, For pitch internal parameters, As a candidate pitch external reference, This is the new pitch extrinsic parameter obtained after updating the candidate pitch extrinsic parameters based on the pitch intrinsic parameter value.

[0175] Combining formulas 6 and 7, it can be seen that by modifying the pitch intrinsic parameter values... This achieves the purpose of modifying the candidate pitch extrinsic parameters. When the candidate pitch extrinsic parameters are modified, the position (x, y, z) of the data points is recalculated, thereby updating the point cloud.

[0176] As mentioned earlier, there are installation errors in the galvanometer and the expanding lens, which cause the ground height obtained by the lidar based on different pitch angles to be inconsistent. The embodiments of this application determine the target pitch intrinsic parameters based on multiple frames of point cloud. The target pitch intrinsic parameters can correct the point cloud distortion that appears in the pitch angle due to installation errors, so that the lidar can obtain a consistent ground height based on each frame of point cloud.

[0177] In this embodiment, the target pitch intrinsic parameters are determined through steps S71 to S74, as follows:

[0178] Step S71: Configure multiple pitch intrinsic parameter values ​​for the point cloud in the same frame.

[0179] In this application embodiment, multiple pitch intrinsic parameter values ​​are configured for the same frame of point cloud. For example, in this application embodiment, a first pitch intrinsic parameter value u1 and a second pitch intrinsic parameter value u2 are configured for the left edge point cloud, where u1 can be 0 and u2 can be r.

[0180] Step S72: Update the point cloud in the same frame using multiple pitch intrinsic parameter values ​​to obtain multiple ground heights of the point cloud in the same frame.

[0181] In this embodiment, each data point in the same frame point cloud is updated using Formula 7 by combining each pitch intrinsic parameter value, thereby completing the update of the same frame point cloud. It can be understood that when the first pitch intrinsic parameter value u1=0, the same frame point cloud remains unchanged.

[0182] Multiple pitch intrinsic values ​​include the first pitch intrinsic value and the second pitch intrinsic value, and the point cloud in the same frame includes the left edge point cloud, the right edge point cloud and the middle point cloud.

[0183] In this embodiment, the left edge point cloud is updated using the first pitch intrinsic parameter value to obtain the fifth point cloud, and the left edge point cloud is updated using the second pitch intrinsic parameter value to obtain the sixth point cloud.

[0184] In this embodiment, the right edge point cloud is updated using the first pitch intrinsic parameter value to obtain the seventh point cloud, and the right edge point cloud is updated using the second pitch intrinsic parameter value to obtain the eighth point cloud.

[0185] In this embodiment, the intermediate point cloud is updated using the first pitch intrinsic parameter value to obtain the ninth point cloud, and the intermediate point cloud is updated using the second pitch intrinsic parameter value to obtain the tenth point cloud.

[0186] In this embodiment, the ground height z11 is determined based on the fifth point cloud, z21 is determined based on the sixth point cloud, z12 is determined based on the seventh point cloud, z22 is determined based on the eighth point cloud, z13 is determined based on the ninth point cloud, and z13 is determined based on the tenth point cloud.

[0187] Step S73: Based on multiple pitch intrinsic parameter values ​​of the point cloud in the same frame and the corresponding ground height, obtain the correlation between the pitch intrinsic parameter values ​​of the point cloud in the same frame and the ground height.

[0188] In this embodiment, multiple pitch intrinsic parameter values ​​of the same frame point cloud are linearly fitted with the corresponding ground height to obtain a second fitted straight line corresponding to the same frame point cloud.

[0189] In this embodiment, a second fitted straight line L21 is generated based on the first pitch intrinsic parameter value u1 and the ground height z11, and the second pitch intrinsic parameter value u2 and the ground height z21.

[0190] In this embodiment, a second fitted straight line L22 is generated based on the first pitch intrinsic value u1 and the ground height z12 and the second pitch intrinsic value u2 and the ground height z22.

[0191] In this embodiment, a second fitted straight line L23 is generated based on the first pitch intrinsic value u1 and the ground height z13, and the second pitch intrinsic value u2 and the ground height z23.

[0192] The second fitted line L21 describes the relationship between the pitch intrinsic value and ground height of the left edge point cloud under different pitch intrinsic value settings. The second fitted line L22 describes the relationship between the pitch intrinsic value and ground height of the right edge point cloud under different pitch intrinsic value settings. The second fitted line L23 describes the relationship between the pitch intrinsic value and ground height of the middle point cloud under different pitch intrinsic value settings.

[0193] The second fitted lines L21, L22, and L23 are all straight lines with pitch intrinsic parameters as independent variables and ground height as dependent variables.

[0194] The second fitted line L21 is represented by a third linear function, which is calculated as follows: the third linear function y=k3x+b3 is constructed in advance. The fifth fitting point (u1,z11) and the sixth fitting point (u2,z21) are substituted into the third linear function y=k3x+b3 to obtain k3 and b3, and then the expression of the third linear function is obtained.

[0195] The second fitted line L22 is represented by the fourth linear function, which is calculated as follows: the fourth linear function y=k4x+b4 is constructed in advance. The seventh fitting point (u1,z12) and the eighth fitting point (u2,z22) are substituted into the fourth linear function y=k4x+b4 to obtain k4 and b4, and then the expression of the fourth linear function is obtained.

[0196] The second fitted line L23 is represented by the fifth linear function, which is calculated as follows: the fifth linear function y=k5x+b5 is constructed in advance. The ninth fitting point (u1,z13) and the tenth fitting point (u2,z23) are substituted into the fifth linear function y=k5x+b5 to obtain k5 and b5, and then the expression of the fifth linear function is obtained.

[0197] Step S74: Determine the target pitch intrinsic parameters based on the correlation between the point clouds in each frame.

[0198] Each frame of point cloud corresponds to a first pitch line, a second pitch line, and a third pitch line, all of which are second fitted lines. For example, the first pitch line is the second fitted line L21 mentioned above, the second pitch line is the second fitted line L22, and the third pitch line is the second fitted line L23. In this embodiment, steps S741 to S747 determine the target pitch intrinsic parameters based on the correlation between the point clouds in each frame, as detailed below:

[0199] Step S741: Obtain the reference pitch intrinsic parameters.

[0200] Step S742: Determine the first ground height based on the first pitch line and the reference pitch intrinsic parameter.

[0201] Step S743: Determine the second ground height based on the second pitch line and the reference pitch intrinsic parameter.

[0202] Step S744: Determine the third ground height based on the third pitch line and the reference pitch intrinsic parameters.

[0203] Step S745: Calculate the first absolute value of the difference between the first ground height and the second ground height.

[0204] Step S746: Calculate the second absolute value of the difference between the second ground height and the third ground height.

[0205] Step S747: In response to the first absolute value being equal to the second absolute value, determine the reference pitch intrinsic value as the target pitch intrinsic value.

[0206] In step S741, for example, in this embodiment of the application, the reference pitch intrinsic parameter is set to e.

[0207] In step S742, for example, the reference pitch intrinsic parameter e is substituted into the third linear function y=k3x+b3 of the second fitted line L21 to obtain the first ground height y1.

[0208] In step S743, for example, the reference pitch intrinsic parameter e is substituted into the fourth linear function y=k4x+b4 of the second fitted line L22 to obtain the second ground height y2.

[0209] In step S744, for example, the reference pitch intrinsic parameter e is substituted into the fifth linear function y=k5x+b5 of the second fitted line L23 to obtain the third ground height y3.

[0210] In step S745, for example, the first absolute value of the difference between the first ground height y1 and the second ground height y2 is... .

[0211] In step S746, for example, the second absolute value of the difference between the second ground height y2 and the third ground height y3 is... .

[0212] In step S747, when the first absolute value Equal to the second absolute value This indicates that when the lidar observes the ground height at different elevation angles, it obtains the same height result. In this case, the reference elevation intrinsic parameter can correct the point cloud distortion caused by installation errors in the elevation angle direction. Therefore, in this embodiment, the reference elevation intrinsic parameter is used as the target elevation intrinsic parameter. When the first absolute value... Not equal to the second absolute value When the lidar observes the ground height at different pitch angles, it obtains different height results. At this time, the reference pitch intrinsic parameter cannot correct the point cloud distortion caused by the installation error in the pitch angle direction. Therefore, the embodiment of this application needs to update the reference pitch intrinsic parameter e and continue to search for a reference pitch intrinsic parameter e that can meet the requirements of step S747.

[0213] 6.3) Determine the target roll internal parameters

[0214] The embodiments of this application determine the target roll intrinsic parameters through the following steps, as shown below: Based on multi-frame point clouds, determine the target roll intrinsic parameters.

[0215] The target roll intrinsic parameter is a roll intrinsic parameter value that enables the point cloud in each frame to obtain a consistent ground height. The roll intrinsic parameter value is the increment of the roll angle corresponding to the data point in the point cloud. In this embodiment, the target roll intrinsic parameter is determined based on multiple frames of point cloud through steps S81 to S84, as shown below:

[0216] Step S81: Configure multiple roll intrinsic parameter values ​​for the point cloud in the same frame.

[0217] In this application embodiment, multiple roll intrinsic parameter values ​​are configured for the same frame of point cloud. For example, in this application embodiment, a first roll intrinsic parameter value v1 and a second roll intrinsic parameter value v2 are configured for the left edge point cloud, where v1 can be 0 and v2 can be q.

[0218] Step S82: Update the point cloud in the same frame using multiple roll intrinsic parameter values ​​to obtain multiple ground heights of the point cloud in the same frame.

[0219] Multiple roll intrinsic values ​​include a first roll intrinsic value and a second roll intrinsic value, and the point cloud in the same frame includes the left edge point cloud and the right edge point cloud.

[0220] Please refer to the following: Figure 12 and Figure 13 In this embodiment, the left edge point cloud is updated using a first roll intrinsic value to obtain the eleventh point cloud, and the left edge point cloud is updated using a second roll intrinsic value to obtain the twelfth point cloud. In this embodiment, the right edge point cloud is updated using a first roll intrinsic value to obtain the thirteenth point cloud, and the right edge point cloud is updated using a second roll intrinsic value to obtain the fourteenth point cloud. In this embodiment, the ground height a11 is determined based on the eleventh point cloud, the ground height a21 is determined based on the twelfth point cloud, the ground height a12 is determined based on the thirteenth point cloud, and the ground height a22 is determined based on the fourteenth point cloud.

[0221] Step S83: Based on multiple roll intrinsic parameter values ​​of the point cloud in the same frame and the corresponding ground height, obtain the correlation between the roll intrinsic parameter values ​​of the point cloud in the same frame and the ground height.

[0222] This application embodiment performs linear fitting on multiple roll intrinsic parameter values ​​of the same frame point cloud with the corresponding ground height to obtain a third fitted straight line corresponding to the same frame point cloud. Specifically, this application embodiment generates a third fitted straight line L31 based on the first roll intrinsic parameter value v1 and ground height a11 and the second roll intrinsic parameter value v2 and ground height a21, and generates a third fitted straight line L32 based on the first roll intrinsic parameter value v1 and ground height a12 and the second roll intrinsic parameter value v2 and ground height a22.

[0223] The third fitted line L31 describes the relationship between the roll intrinsic value and the ground height of the left edge point cloud under different roll intrinsic value settings. The third fitted line L32 describes the relationship between the roll intrinsic value and the ground height of the right edge point cloud under different roll intrinsic value settings.

[0224] Both the third fitted line L31 and the third fitted line L32 are straight lines with the roll intrinsic parameter as the independent variable and the ground height as the dependent variable.

[0225] The third fitted line L31 is represented by the sixth linear function, which is calculated as follows: the sixth linear function y=k6x+b6 is constructed in advance. The eleventh fitted point (v1,a11) and the twelfth fitted point (v2,a21) are substituted into the above sixth linear function y=k1x+b1 to obtain k6 and b6, and then the expression of the sixth linear function is obtained.

[0226] The third fitted line L32 is represented by the seventh linear function, which is calculated as follows: the seventh linear function y=k7x+b7 is constructed in advance. The thirteenth fitted point (v1,a12) and the fourteenth fitted point (v2,a22) are substituted into the seventh linear function y=k7x+b7 to obtain k7 and b7, and then the expression of the seventh linear function is obtained.

[0227] Step S84: Determine the target roll intrinsic parameters based on the correlation between the point clouds in each frame.

[0228] In this embodiment, the roll intrinsic value corresponding to the intersection point of the third fitted line corresponding to each point cloud is determined as the target roll intrinsic value. Specifically, in this embodiment, the roll intrinsic value corresponding to the intersection point of the third fitted line L31 and the third fitted line L32 is determined as the target roll intrinsic value.

[0229] The target roll intrinsic parameter enables the lidar to obtain the same distance result when observing ground height at different roll angles. Therefore, the target roll intrinsic parameter corrects the point cloud distortion caused by installation errors in the roll angle direction. Furthermore, the embodiments of this application can obtain accurate and reliable target roll intrinsic parameters by relying on only a few frames of point cloud (left edge point cloud and right edge point cloud), which is efficient and accurate.

[0230] 7. Update the candidate extrinsic data using the target intrinsic parameter data to obtain the target extrinsic parameter data.

[0231] In this embodiment, candidate extrinsic data are updated based on target intrinsic data to obtain target extrinsic data, which includes target yaw extrinsic data, target pitch extrinsic data, and target roll extrinsic data.

[0232] After the operation in point six, the embodiment of this application has obtained the target intrinsic parameter data, which will align the angle between the galvanometer and the expanding lens with the actual physical situation. However, when the lidar uses the target intrinsic parameter data and candidate extrinsic parameter data to calculate the position of each data point in each frame of the point cloud, point cloud offset will still occur. Therefore, the embodiment of this application needs to use the target intrinsic parameter data to update the candidate extrinsic parameter data and recalibrate the candidate yaw extrinsic parameter, candidate pitch extrinsic parameter, and candidate roll extrinsic parameter, so as to obtain more accurate target yaw extrinsic parameter, target pitch extrinsic parameter, and target roll extrinsic parameter respectively, so that the point cloud no longer shifts or becomes distorted.

[0233] The embodiments of this application use the following steps to update the candidate extrinsic parameter data, as follows: subtract the candidate yaw extrinsic parameter from the target yaw intrinsic parameter by a preset multiple to obtain the target yaw extrinsic parameter; subtract the candidate pitch extrinsic parameter from the target pitch intrinsic parameter by a preset multiple to obtain the target pitch extrinsic parameter; and subtract the candidate roll extrinsic parameter from the target roll intrinsic parameter by a preset multiple to obtain the target roll extrinsic parameter.

[0234] The preset magnification is the magnification of the wide-angle lens. For example, if the preset magnification is 2, then:

[0235] roll_new=roll-2*Inter_Roll

[0236] pitch_new=pitch-2*Inter_Pitch

[0237] yaw_new=yaw-2*Inter_Yaw

[0238] Where roll_new is the target roll extrinsic parameter, pitch_new is the target pitch extrinsic parameter, yaw_new is the target yaw extrinsic parameter, roll is the candidate roll extrinsic parameter, pitch is the candidate pitch extrinsic parameter, and yaw is the candidate yaw extrinsic parameter.

[0239] 8. Verify whether the target intrinsic parameter data meets the requirements.

[0240] In this embodiment, the target yaw, pitch, and roll intrinsic parameters are written into the lidar, and the lidar is controlled to re-acquire point clouds. Following the method described in the previous embodiment, the tested inter-plate distance value and the tested ground height are obtained respectively. When the tested inter-plate distance value is less than a first threshold and the tested ground height is less than a second threshold, this embodiment generates calibration success information and saves the target yaw, pitch, and roll intrinsic parameters. When the tested inter-plate distance value is greater than the first threshold, or the tested ground height is greater than the second threshold, this embodiment generates calibration failure information and re-determines the target yaw, pitch, and roll intrinsic parameters.

[0241] In summary, the embodiments of this application have at least the following technical effects:

[0242] 1) The embodiments of this application correct the point cloud distortion in the yaw angle caused by installation errors by calibrating the target yaw intrinsic parameters, which is beneficial for the lidar to obtain reliable and accurate point clouds.

[0243] 2) The embodiments of this application calibrate the target pitch intrinsic parameters and correct the point cloud distortion that appears in the pitch angle due to installation errors, which is beneficial for the lidar to obtain reliable and accurate point clouds.

[0244] 3) The embodiments of this application correct the point cloud distortion in the roll angle caused by installation errors by calibrating the target roll intrinsic parameters, which is beneficial for the lidar to obtain reliable and accurate point clouds.

[0245] 4) Based on points 1) to 3), the embodiments of this application can correct the point cloud distortion that occurs in the lidar at any posture due to installation errors, which is beneficial for the lidar to obtain reliable and accurate point clouds.

[0246] 5) The embodiments of this application update the candidate extrinsic data based on the target intrinsic data, thereby obtaining more accurate and reliable target extrinsic data. This enables the lidar to correct the point cloud distortion based on the target extrinsic data and obtain a more accurate and reliable point cloud.

[0247] 6) The internal and external parameter calibration method provided in this application embodiment can be applied to LiDAR, providing the industry with a new approach to calibrate the wide-angle lens even in point cloud scenarios with sparse pixels.

[0248] 7) The embodiments of this application can automatically control the turntable to drive the LiDAR to rotate, automatically process the point cloud collected by the LiDAR, and automatically calibrate the internal and external parameters. The entire calibration process is an automated process, which can reduce the error of manual calibration and improve the calibration efficiency, accuracy and stability.

[0249] It should be noted that in the above embodiments, there is no necessarily a certain order between the steps. Those skilled in the art can understand from the description of the embodiments of this application that the above steps may have different execution orders in different embodiments, that is, they may be executed in parallel or in turn, etc.

[0250] As another aspect of the embodiments of this application, this application provides an intrinsic parameter calibration device for a lidar. The lidar's intrinsic parameter calibrator can access the memory, invoke instructions for execution, and complete the lidar intrinsic parameter calibration method described in the various embodiments above.

[0251] In some implementations, the intrinsic parameter calibration device of the lidar can also be built from hardware components. For example, the intrinsic parameter calibration device of the lidar can be built from one or more chips, and the chips can work together to complete the intrinsic parameter calibration method of the lidar described in the above implementations. As another example, the intrinsic parameter calibration device of the lidar can also be built from various logic devices, such as general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), microcontrollers, ARM (Acorn RISC Machine) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components.

[0252] Please see Figure 14 The lidar intrinsic parameter calibration device 140 includes a yaw intrinsic parameter calibration module 141. The yaw intrinsic parameter calibration module 141 is used to acquire multiple frames of point clouds obtained by the lidar scanning the first calibration board and the second calibration board in different attitudes. Based on the multiple frames of point clouds, the target yaw intrinsic parameter is determined. The target yaw intrinsic parameter is a yaw intrinsic parameter value that enables the point cloud in each frame to obtain a consistent inter-board distance value. The yaw intrinsic parameter value is the increment of the yaw angle corresponding to the data point in the point cloud. The inter-board distance value is the distance from the first calibration board to the second calibration board.

[0253] In some embodiments, the yaw intrinsic parameter calibration module 141 is specifically used to: configure multiple yaw intrinsic parameter values ​​for the same frame point cloud, update the same frame point cloud with the multiple yaw intrinsic parameter values ​​respectively to obtain multiple inter-board distance values ​​of the same frame point cloud, obtain the correlation relationship between the yaw intrinsic parameter values ​​and the inter-board distance values ​​corresponding to the same frame point cloud based on the multiple yaw intrinsic parameter values ​​and the corresponding inter-board distance values ​​of the same frame point cloud, and determine the target yaw intrinsic parameter according to the correlation relationship corresponding to the point cloud in each frame.

[0254] In some embodiments, the yaw intrinsic parameter calibration module 141 is further specifically used to: perform linear fitting on multiple yaw intrinsic parameter values ​​of the same frame point cloud and the corresponding inter-plate distance values ​​to obtain the first fitted straight line corresponding to the same frame point cloud.

[0255] In some embodiments, the yaw intrinsic parameter calibration module 141 is further specifically used to: determine the yaw intrinsic parameter value corresponding to the intersection point of the first fitted straight line corresponding to each frame of the point cloud as the target yaw intrinsic parameter.

[0256] Please continue reading. Figure 14The lidar intrinsic parameter calibration device 140 further includes a pitch intrinsic parameter calibration module 142. The pitch intrinsic parameter calibration module 142 is used to determine the target pitch intrinsic parameter based on multiple frames of the point cloud. The target pitch intrinsic parameter is a pitch intrinsic parameter value that makes the first absolute difference between the ground height of the first target point cloud and the ground height of the third target point cloud equal to the second absolute difference between the ground height of the second target point cloud and the ground height of the third target point cloud. The first field of view corresponding to the first target point cloud and the second field of view corresponding to the second target point cloud are symmetrical about the third field of view corresponding to the third target point cloud. The pitch intrinsic parameter value is the increment of the pitch angle corresponding to the data point in the point cloud.

[0257] In some embodiments, the pitch intrinsic parameter calibration module 142 is specifically used to: configure multiple pitch intrinsic parameter values ​​for the same frame point cloud, update the same frame point cloud with the multiple pitch intrinsic parameter values ​​respectively to obtain multiple ground heights of the same frame point cloud, obtain the correlation between the pitch intrinsic parameter values ​​and ground heights corresponding to the same frame point cloud based on the multiple pitch intrinsic parameter values ​​of the same frame point cloud and the corresponding ground heights, and determine the target pitch intrinsic parameter according to the correlation between the point cloud in each frame.

[0258] In some embodiments, the pitch intrinsic parameter calibration module 142 is specifically used to: linearly fit multiple pitch intrinsic parameter values ​​of the same frame point cloud with the corresponding ground height to obtain a second fitted straight line corresponding to the same frame point cloud.

[0259] In some embodiments, each frame of the point cloud corresponds to a first pitch line, a second pitch line, and a third pitch line. The first pitch line, the second pitch line, and the third pitch line are all second fitted lines. The pitch intrinsic parameter calibration module 142 is specifically used to: obtain a reference pitch intrinsic parameter; determine a first ground height based on the first pitch line and the reference pitch intrinsic parameter; determine a second ground height based on the second pitch line and the reference pitch intrinsic parameter; determine a third ground height based on the third pitch line and the reference pitch intrinsic parameter; calculate a first absolute value of the difference between the first ground height and the second ground height; calculate a second absolute value of the difference between the second ground height and the third ground height; and determine the reference pitch intrinsic parameter as the target pitch intrinsic parameter in response to the first absolute value being equal to the second absolute value.

[0260] In some embodiments, please continue reading Figure 14 The lidar intrinsic parameter calibration device 140 also includes a roll intrinsic parameter calibration module 143, which is used to determine the target roll intrinsic parameter based on multiple frames of point cloud. The target roll intrinsic parameter is a roll intrinsic parameter value that enables the point cloud in each frame to obtain a consistent ground height. The roll intrinsic parameter value is the increment of the roll angle corresponding to the data point in the point cloud.

[0261] In some embodiments, the roll intrinsic parameter calibration module 143 is specifically used to: configure multiple roll intrinsic parameter values ​​for the same frame point cloud, update the same frame point cloud with the multiple roll intrinsic parameter values ​​respectively to obtain multiple ground heights of the same frame point cloud, obtain the correlation between the roll intrinsic parameter values ​​and ground heights corresponding to the same frame point cloud based on the multiple roll intrinsic parameter values ​​and corresponding ground heights of the same frame point cloud, and determine the target roll intrinsic parameter according to the correlation between the point cloud in each frame.

[0262] In some embodiments, the roll intrinsic parameter calibration module 143 is specifically used to: linearly fit multiple roll intrinsic parameter values ​​of the same frame point cloud with the corresponding ground height to obtain a third fitted straight line corresponding to the same frame point cloud.

[0263] In some embodiments, the roll intrinsic parameter calibration module 143 is specifically used to: determine the roll intrinsic parameter value corresponding to the intersection point of the third fitted line corresponding to each point cloud as the target roll intrinsic parameter.

[0264] In some embodiments, the field of view range of the lidar is [m, n]. The yaw intrinsic parameter calibration module 141 is further specifically used to: acquire the point cloud obtained by the lidar scanning the first calibration plate and the second calibration plate in a first attitude, wherein the first attitude is the attitude when the laser line of the lidar at the m-th field of view hits the first calibration plate; acquire the point cloud obtained by the lidar scanning the first calibration plate and the second calibration plate in a second attitude, wherein the second attitude is the attitude when the laser line of the lidar at the n-th field of view hits the second calibration plate; and acquire the point cloud obtained by the lidar scanning the first calibration plate and the second calibration plate in a third attitude, wherein the third attitude is the attitude when the laser line of the lidar at the n-th field of view hits the second calibration plate. The posture of the laser line in the field of view when it hits the middle of the first calibration plate and the second calibration plate.

[0265] As another aspect of this application, this application provides a method for calibrating the intrinsic and extrinsic parameters of a lidar. Please refer to... Figure 15 The lidar's intrinsic and extrinsic parameter calibration device 150 includes an extrinsic parameter calibration module 151 and an intrinsic parameter calibration module 152. The extrinsic parameter calibration module 151 is used to determine candidate extrinsic parameter data for the lidar, including candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters. The intrinsic parameter calibration module 152 is used to acquire target intrinsic parameter data obtained based on the aforementioned intrinsic parameter calibration method, including target yaw intrinsic parameters, target pitch intrinsic parameters, and target roll intrinsic parameters.

[0266] In some embodiments, the extrinsic parameter calibration module 151 is specifically used to: acquire true ground data, the true ground data being used to represent the true ground of the calibration site, the first calibration plate and the second calibration plate being disposed at intervals on the calibration site, determine a first normal vector of the true ground based on the true ground data, determine a second normal vector of the actual ground based on the point cloud, the actual ground being the ground of the calibration site detected by the lidar, and determine a first angle between the first normal vector and the second normal vector, the first angle being the candidate yaw extrinsic parameter.

[0267] In some embodiments, the extrinsic parameter calibration module 151 is specifically used to: acquire truth panel data, the truth panel data being used to represent the truth panel corresponding to the first calibration panel and the second calibration panel; determine the third normal vector of the truth panel based on the truth panel data; determine the fourth normal vector of the actual panel based on the point cloud, the actual panel being the panel obtained by the lidar detecting the first calibration panel or the second calibration panel; and determine the second included angle between the third normal vector and the fourth normal vector, the second included angle being the candidate pitch extrinsic parameter.

[0268] In some embodiments, the extrinsic parameter calibration module 151 is specifically used to: acquire true ground data and true plate data, wherein the true ground data represents the true ground of the calibration site, and the true plate data represents the true plate surface corresponding to the first calibration plate and the second calibration plate; determine a first normal vector of the true ground based on the true ground data; determine a third normal vector of the true plate surface based on the true plate data; perform a cross product of the first normal vector and the third normal vector to obtain a fifth normal vector; determine a second normal vector of the actual ground and a fourth normal vector of the first calibration plate or the second calibration plate based on the point cloud, wherein the actual ground is the ground of the calibration site detected by the lidar; perform a cross product of the second normal vector and the fourth normal vector to obtain a sixth normal vector; and determine a third included angle between the fifth normal vector and the sixth normal vector, wherein the third included angle is the candidate roll extrinsic parameter.

[0269] In some embodiments, the extrinsic parameter calibration module 151 is specifically used to: update the candidate extrinsic parameter data based on the target intrinsic parameter data to obtain target extrinsic parameter data, wherein the target extrinsic parameter data includes target yaw extrinsic parameter, target pitch extrinsic parameter and target roll extrinsic parameter.

[0270] In some embodiments, the extrinsic parameter calibration module 151 is specifically used to: subtract the target yaw intrinsic parameter from the candidate yaw extrinsic parameter by a preset multiple to obtain the target yaw extrinsic parameter; subtract the target pitch intrinsic parameter from the candidate pitch extrinsic parameter by a preset multiple to obtain the target pitch extrinsic parameter; and subtract the target roll intrinsic parameter from the candidate roll extrinsic parameter by a preset multiple to obtain the target roll extrinsic parameter.

[0271] It should be noted that the aforementioned lidar internal parameter calibration device or lidar internal / external parameter calibration device can execute the lidar internal parameter calibration method or lidar internal / external parameter calibration method provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects of the execution method. Technical details not described in detail in the embodiments of the lidar internal parameter calibration device or lidar internal / external parameter calibration device can be found in the lidar internal parameter calibration method or lidar internal / external parameter calibration method provided in the embodiments of this application.

[0272] See Figure 16 , Figure 16 This is a schematic diagram of a lidar system provided in an embodiment of this application. The lidar 160 includes one or more processors 161 and a memory 162. The memory 162 is connected to one or more processors 161, for example, via a bus.

[0273] Processor 161 is configured to support the lidar in performing the corresponding functions in the methods described in the above-described method embodiments. The processor may be a central processing unit (CPU), a network processor (NP), a hardware chip, or any combination thereof. The aforementioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), generic array logic (GAL), or any combination thereof.

[0274] Memory 162 is used to store program code, etc. Memory may include volatile memory (VM), such as random access memory (RAM); memory may also include non-volatile memory (NVM), such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); memory may also include combinations of the above types of memory.

[0275] The memory 162 can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the lidar intrinsic parameter calibration method or lidar intrinsic and extrinsic parameter calibration method in the embodiments of this application. The processor executes the non-volatile software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the lidar intrinsic parameter calibration method or lidar intrinsic and extrinsic parameter calibration method and lidar intrinsic and extrinsic parameter calibration device, thereby realizing the functions of each module or unit of the lidar intrinsic parameter calibration method or lidar intrinsic and extrinsic parameter calibration method and lidar intrinsic and extrinsic parameter calibration device provided in the above method embodiments.

[0276] The memory may include a program storage area and a data storage area, wherein the program storage area may store the operating system and application programs required for at least one function. The data storage area may store data created based on the use of the lidar's intrinsic parameter calibration device or the lidar's extrinsic and extra-extrinsic parameter calibration device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the lidar's intrinsic parameter calibration device or the lidar's extrinsic and extra-extrinsic parameter calibration device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0277] The one or more modules are stored in the memory. When executed by the one or more processors, they perform the intrinsic parameter calibration method of the lidar or the intrinsic and extrinsic parameter calibration method of the lidar in any of the above method embodiments. For example, they perform the method steps described in the above method embodiments to realize the functions of the modules described in the above device embodiments.

[0278] This application also provides a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the method described in the foregoing embodiments.

[0279] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0280] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A method for calibrating the intrinsic parameters of a lidar, characterized in that, include: Multiple point clouds were obtained by scanning the first and second calibration boards with the lidar in different postures; Based on multiple frames of the point cloud, a target yaw intrinsic parameter is determined. The target yaw intrinsic parameter is a yaw intrinsic parameter value that enables the point cloud in each frame to obtain a consistent inter-plate distance value. The yaw intrinsic parameter value is the increment of the yaw angle corresponding to the data point in the point cloud, and the inter-plate distance value is the distance from the first calibration plate to the second calibration plate. Determining the target yaw intrinsic parameter based on multiple frames of the point cloud includes: configuring multiple yaw intrinsic parameter values ​​for the same frame of the point cloud; updating the same frame of the point cloud with the multiple yaw intrinsic parameter values ​​respectively to obtain multiple inter-plate distance values ​​for the same frame of the point cloud; obtaining the correlation between the yaw intrinsic parameter values ​​and the corresponding inter-plate distance values ​​for the same frame of the point cloud based on the multiple yaw intrinsic parameter values ​​and the corresponding inter-plate distance values; and determining the target yaw intrinsic parameter based on the correlation between the point cloud in each frame.

2. The method according to claim 1, characterized in that, The process of obtaining the correlation between the yaw intrinsic parameter values ​​and the corresponding inter-plate distance values ​​of the same frame point cloud based on multiple yaw intrinsic parameter values ​​and the corresponding inter-plate distance values ​​includes: Linear fitting is performed on multiple yaw intrinsic parameter values ​​of the same frame point cloud and the corresponding inter-board distance values ​​to obtain the first fitted straight line corresponding to the same frame point cloud.

3. The method according to claim 2, characterized in that, The step of determining the target yaw intrinsic parameters based on the correlation between the multiple point clouds includes: The yaw intrinsic parameter value corresponding to the intersection point of the first fitted straight line corresponding to the point cloud in each frame is determined as the target yaw intrinsic parameter.

4. The method according to claim 1, characterized in that, Also includes: Based on the point cloud of multiple frames, target pitch intrinsic parameters are determined. The target pitch intrinsic parameters are pitch intrinsic parameter values ​​that make the first absolute difference between the ground height of the first target point cloud and the ground height of the third target point cloud equal to the second absolute difference between the ground height of the second target point cloud and the ground height of the third target point cloud. The first field of view corresponding to the first target point cloud and the second field of view corresponding to the second target point cloud are symmetrical about the third field of view corresponding to the third target point cloud. The pitch intrinsic parameter value is the increment of the pitch angle corresponding to the data point in the point cloud.

5. The method according to claim 4, characterized in that, The determination of target pitch intrinsic parameters based on multiple frames of the point cloud includes: Configure multiple pitch intrinsic parameter values ​​for the point cloud in the same frame; The same frame point cloud is updated using multiple pitch intrinsic parameter values ​​to obtain multiple ground heights of the same frame point cloud; Based on multiple pitch intrinsic parameter values ​​of the same frame point cloud and their corresponding ground heights, the correlation between the pitch intrinsic parameter values ​​of the same frame point cloud and the ground heights is obtained; and Based on the correlation between the point clouds in each frame, the target pitch intrinsic parameters are determined.

6. The method according to claim 5, characterized in that, The process of obtaining the correlation between the pitch intrinsic values ​​of the same frame point cloud and the corresponding ground height based on multiple pitch intrinsic values ​​and the corresponding ground height includes: Multiple pitch intrinsic parameter values ​​of the same frame point cloud are linearly fitted with the corresponding ground height to obtain a second fitted straight line corresponding to the same frame point cloud.

7. The method according to claim 6, characterized in that, Each frame of the point cloud corresponds to a first pitch line, a second pitch line, and a third pitch line. The first pitch line, the second pitch line, and the third pitch line are all second fitted lines. The step of determining the target pitch intrinsic parameters based on the correlation between the point clouds in each frame includes: Obtain reference pitch parameters; The first ground height is determined based on the first pitch line and the reference pitch intrinsic parameter; The second ground height is determined based on the second pitch line and the reference pitch intrinsic parameter; The third ground height is determined based on the third pitch line and the reference pitch intrinsic parameter; Calculate the first absolute value of the difference between the first ground height and the second ground height; Calculate the second absolute value of the difference between the second ground height and the third ground height; In response to the first absolute value being equal to the second absolute value, the reference pitch intrinsic value is determined to be the target pitch intrinsic value.

8. The method according to claim 1, characterized in that, Also includes: Based on the point cloud in multiple frames, a target roll intrinsic parameter is determined, wherein the target roll intrinsic parameter is a roll intrinsic parameter value that enables the point cloud in each frame to obtain a consistent ground height, and the roll intrinsic parameter value is the increment of the roll angle corresponding to the data point in the point cloud.

9. The method according to claim 8, characterized in that, The determination of target roll intrinsic parameters based on multiple frames of the point cloud includes: Configure multiple roll intrinsic parameter values ​​for the point cloud in the same frame; The same frame point cloud is updated using multiple roll intrinsic parameter values ​​to obtain multiple ground heights of the same frame point cloud; Based on multiple roll intrinsic parameter values ​​of the same frame point cloud and their corresponding ground heights, the correlation between the roll intrinsic parameter values ​​of the same frame point cloud and the ground heights is obtained; and Based on the correlation between the point clouds in each frame, the target roll intrinsic parameters are determined.

10. The method according to claim 9, characterized in that, The process of obtaining the correlation between the roll intrinsic values ​​of the same frame point cloud and the corresponding ground height based on multiple roll intrinsic values ​​and the ground height includes: Multiple roll intrinsic parameter values ​​of the same frame point cloud are linearly fitted with the corresponding ground height to obtain the third fitted straight line corresponding to the same frame point cloud.

11. The method according to claim 10, characterized in that, The step of determining the target roll intrinsic parameters based on the correlation relationship between the point clouds in each frame includes: The roll intrinsic value corresponding to the intersection point of the third fitted line corresponding to each point cloud is determined as the target roll intrinsic value.

12. The method according to any one of claims 1 to 11, characterized in that, The field of view of the lidar is [m, n]. The acquisition of multiple point clouds obtained by the lidar scanning the first and second calibration boards in different postures includes: The point cloud obtained by the lidar scanning the first calibration plate and the second calibration plate in the first posture is the posture when the laser line of the mth field of view of the lidar hits the first calibration plate. The point cloud obtained by the lidar scanning the first calibration plate and the second calibration plate in the second posture is the posture when the laser line of the nth field of view of the lidar hits the second calibration plate. The point cloud obtained by the lidar scanning the first and second calibration boards in a third posture is acquired, wherein the third posture is the first position of the lidar. The posture of the laser line in the field of view when it hits the middle of the first calibration plate and the second calibration plate.

13. A method for calibrating the intrinsic and extrinsic parameters of a lidar, characterized in that, include: Determine the candidate extrinsic parameters of the lidar, including candidate yaw extrinsic parameters, candidate pitch extrinsic parameters, and candidate roll extrinsic parameters; Obtain target intrinsic parameter data based on the intrinsic parameter calibration method as described in any one of claims 1 to 12, wherein the target intrinsic parameter data includes target yaw intrinsic parameter, target pitch intrinsic parameter and target roll intrinsic parameter.

14. The method according to claim 13, characterized in that, The candidate extrinsic parameter data includes candidate yaw extrinsic parameters, and determining the candidate extrinsic parameter data of the lidar includes: Obtain true ground data, which is used to represent the true ground of the calibration site. The first calibration plate and the second calibration plate are spaced apart on the calibration site. Determine the first normal vector of the true ground based on the true ground data; The second normal vector of the actual ground is determined based on the point cloud, wherein the actual ground is the ground of the calibration site detected by the lidar; Determine the first angle between the first normal vector and the second normal vector, where the first angle is the candidate yaw extrinsic parameter.

15. The method according to claim 13, characterized in that, The candidate extrinsic parameter data includes candidate elevation extrinsic parameters, and determining the candidate extrinsic parameter data of the lidar includes: Obtain truth panel data, which is used to represent the truth panel corresponding to the first calibration board and the second calibration board; The third normal vector of the truth panel is determined based on the truth panel data; The fourth normal vector of the actual board surface is determined based on the point cloud, where the actual board surface is the board surface obtained by the lidar detecting the first calibration board or the second calibration board. Determine the second angle between the third normal vector and the fourth normal vector, whereby the second angle is the candidate pitch extrinsic parameter.

16. The method according to claim 13, characterized in that, The candidate extrinsic parameter data includes candidate roll extrinsic parameters, and determining the candidate extrinsic parameter data of the lidar includes: Acquire ground truth data and ground truth data, wherein the ground truth data is used to represent the ground truth of the calibration site, and the ground truth data is used to represent the ground truth corresponding to the first calibration board and the second calibration board; Determine the first normal vector of the true ground based on the true ground data; The third normal vector of the truth panel is determined based on the truth panel data; The cross product of the first normal vector and the third normal vector is used to obtain the fifth normal vector. The second normal vector of the actual ground and the fourth normal vector of the first calibration plate or the second calibration plate are determined based on the point cloud, wherein the actual ground is the ground of the calibration site detected by the lidar; The cross product of the second normal vector and the fourth normal vector is used to obtain the sixth normal vector. Determine the third angle between the fifth normal vector and the sixth normal vector, whereby the third angle is the candidate roll extrinsic parameter.

17. The method according to claim 13, characterized in that, Also includes: The candidate extrinsic data are updated based on the target intrinsic data to obtain the target extrinsic data, which includes the target yaw extrinsic data, the target pitch extrinsic data, and the target roll extrinsic data.

18. The method according to claim 17, characterized in that, The step of updating the candidate extrinsic data based on the target intrinsic data to obtain the target extrinsic data includes: Subtract the target yaw intrinsic parameter from the candidate yaw extrinsic parameter by a preset multiple to obtain the target yaw extrinsic parameter. Subtract the target pitch intrinsic parameter from the candidate pitch extrinsic parameter by a preset multiple to obtain the target pitch extrinsic parameter; The target roll extrinsic parameter is obtained by subtracting the candidate roll extrinsic parameter from the target roll intrinsic parameter by a preset multiple.

19. A lidar, characterized in that, The device includes a memory and a processor, the memory being connected to the processor, the processor being configured to execute one or more computer programs stored in the memory, the processor causing the lidar to implement the lidar intrinsic parameter calibration method as described in any one of claims 1-12 or the lidar intrinsic and extrinsic parameter calibration method as described in any one of claims 13-18.

20. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions, which, when executed by a processor, cause the processor to perform the intrinsic parameter calibration method for a lidar as described in any one of claims 1-12 or the intrinsic and extrinsic parameter calibration method for a lidar as described in any one of claims 13-18.

Images