Deviation calibration method of laser radar

Abstract

The invention discloses a deviation calibration method of a laser radar. The laser radar is connected with a holder used for driving the laser radar to rotate. The deviation calibration method comprises the steps of controlling a laser radar to perform point cloud scanning on a measurement plane, and obtaining a sampling coordinate value of each sampling point in a laser radar spherical coordinate system and an angle parameter of rotation of the laser radar driven by a holder; according to a conversion relation between the laser radar spherical coordinate system and the holder three-dimensional rectangular coordinate system, a set compensation deviation parameter and an angle parameter corresponding to each sampling point, performing compensation conversion on each sampling coordinate value to obtain a compensation coordinate value of a compensation sampling point; determining a central reference plane of each compensation sampling point; and optimizing an iterative compensation deviation parameter according to the deviation of each compensation sampling point relative to the central reference plane, and taking the compensation deviation parameter when the deviation is minimum as a calibration deviation parameter. The method can achieve the calibration of the compensation amount of the measurement deviation of the laser radar, and improves the measurement accuracy of the laser radar.

Description

Technical Field

[0001] The invention relates to the technical field of laser radar scanning measurement, and in particular to a deviation calibration method of a laser radar. Background Art

[0002] LiDAR is a radar system that emits laser beams to detect the position, speed and other characteristic quantities of a target. Its working principle is to transmit a detection signal (laser beam) to the target, and then compare the received signal reflected from the target (target echo) with the transmitted signal. After appropriate processing, relevant information about the target can be obtained, such as target distance, direction, altitude, speed, attitude, and even shape parameters, thereby detecting, tracking and identifying targets such as aircraft and missiles.

[0003] In order to realize the measurement of scene objects in three-dimensional space by LiDAR, it is necessary to control the LiDAR to collect three-dimensional point clouds of target objects from different angles and directions by using motion scanning. As the use time of LiDAR increases, deformation of the internal LiDAR and the structural components that drive the movement of LiDAR are inevitable, which may lead to a decrease in the measurement accuracy of LiDAR. Summary of the invention

[0004] The purpose of the present invention is to provide a laser radar deviation calibration method, which can calibrate the deviation compensation amount of the laser radar scanning measurement, improve the accuracy of the laser radar in measuring the object scene, and is conducive to the widespread application of the laser radar.

[0005] In order to solve the above technical problems, the present invention provides a deviation calibration method for a laser radar, wherein the laser radar is connected to a pan / tilt for driving the laser radar to rotate; the deviation calibration method comprises:

[0006] Control the laser radar to perform point cloud scanning on the measurement plane, obtain the sampling coordinate value of each sampling point in the laser radar spherical coordinate system, and obtain the angle parameter of the pan-tilt drive the laser radar to rotate corresponding to each sampling point;

[0007] According to the conversion relationship between the laser radar spherical coordinate system and the gimbal three-dimensional rectangular coordinate system, the set compensation deviation parameter and the angle parameter corresponding to each sampling point, each sampling coordinate value is compensated and converted to obtain the compensation coordinate value of the compensation sampling point corresponding to each sampling point in the gimbal three-dimensional rectangular coordinate system;

[0008] Determine the central reference plane of each of the compensation sampling points according to the compensation coordinate value;

[0009] The deviation of each compensation sampling point relative to the central reference plane is taken as an optimization target, and the compensation deviation parameter is iteratively optimized until the compensation deviation parameter corresponding to the minimum deviation is obtained as a calibration deviation parameter.

[0010] In an optional embodiment of the present application, the laser radar includes a plurality of lasers arranged in sequence along a set straight line direction;

[0011] Controlling the laser radar to perform point cloud scanning on the measurement plane includes:

[0012] The laser in the laser radar is controlled to rotate around the first rotation axis, and the laser radar is driven to rotate around the second rotation axis through the gimbal, and each laser is synchronously controlled to perform point cloud scanning on the position points on the measurement plane; wherein the first rotation axis and the set straight line direction are parallel to each other, and the first rotation axis and the second rotation axis are perpendicular to each other.

[0013] In an optional embodiment of the present application, the angle between the scanning direction of each laser and the set straight line direction gradually decreases from the middle position to the two end positions of each laser;

[0014] The second rotation axis is a horizontal rotation axis.

[0015] In an optional embodiment of the present application, the sampling coordinate value includes a distance, a polar angle and an azimuth; the angle parameter includes a roll angle;

[0016] Wherein, the distance is the straight-line distance between the sampling point and the origin of the laser radar spherical coordinate system;

[0017] The polar angle and the azimuth are the angles of the line between the sampling point and the origin of the laser radar spherical coordinate system and the laser radar 3D rectangular coordinate system. Axis and The angle between the axes; wherein the origin of the laser radar three-dimensional rectangular coordinate system coincides with the origin of the laser radar spherical coordinate system; The axis coincides with the first rotation axis, the The axis coincides with the second rotation axis;

[0018] The roll angle is the angle at which the gimbal controls the laser radar to rotate around the second rotation axis when scanning the sampling point.

[0019] In an optional embodiment of the present application, the laser radar spherical coordinate system and the laser radar three-dimensional rectangular coordinate system are two relatively stationary coordinate systems, and both are coordinate systems that move about the second rotation axis relative to the table three-dimensional rectangular coordinate system;

[0020] The X-axis direction and the Y-axis direction of the three-dimensional rectangular coordinate system of the table are both horizontal directions, and the Z-axis direction is the vertical direction; and the Y-axis of the three-dimensional rectangular coordinate system of the table and the three-dimensional rectangular coordinate system of the laser radar are Axes coincide.

[0021] In an optional embodiment of the present application, the compensation deviation parameters set include internal compensation deviation parameters and shaft compensation deviation parameters;

[0022] The internal compensation deviation parameters include a distance compensation amount, a polar angle compensation amount and an azimuth angle compensation amount respectively set for compensating the distance, the polar angle and the azimuth angle;

[0023] The axis compensation deviation parameters include a first translation compensation amount, a second translation compensation amount, a first rotation compensation angle, and a second rotation compensation angle for compensating the roll angle;

[0024] Among them, the first translation compensation amount and the second translation compensation amount are respectively used to compensate for the translation deviations in the Z-axis direction and the X-axis direction of the table top three-dimensional rectangular coordinate system introduced when the table top drives the laser radar to rotate; the first rotation compensation angle and the second rotation compensation angle are respectively used to compensate for the angular deviations relative to the Z-axis and the X-axis of the table top three-dimensional rectangular coordinate system introduced when the table top drives the laser radar to rotate.

[0025] In an optional embodiment of the present application, according to the conversion relationship between the laser radar spherical coordinate system and the gimbal coordinate system and the set compensation deviation parameter, each of the sampled coordinate values ​​is compensated and converted into a compensated coordinate value in the gimbal coordinate system, including:

[0026] The sampled coordinate values ​​are compensated according to the internal compensation deviation parameters and then converted into preliminary compensated coordinate values ​​in the three-dimensional motion rectangular coordinate system of the laser radar. ;in, ; For the The sampling coordinate values ​​of the sampling points; are the distance compensation amount, the polar angle compensation amount and the azimuth angle compensation amount respectively;

[0027] Determine the gimbal compensation matrix according to the conversion relationship between the laser radar three-dimensional rectangular coordinate system and the gimbal three-dimensional rectangular coordinate system and the axis compensation deviation parameter ,in, , ; , , is the first translation compensation amount, is the second translation compensation amount, is the first rotation compensation angle, is the second rotation compensation angle;

[0028] The preliminary compensation coordinate value is converted into the compensation coordinate value according to the pan-tilt compensation matrix and the table conversion matrix ;in, is the table transformation matrix, , For the The roll angle corresponding to each of the sampling points.

[0029] In an optional embodiment of the present application, determining a central reference plane corresponding to the compensation value of each sampling point according to the compensation coordinate value includes:

[0030] A fitting plane corresponding to each of the compensation sampling points is determined according to each of the compensation coordinate values, and the fitting plane is used as the central reference plane.

[0031] In an optional embodiment of the present application, taking the deviation value of each of the compensation coordinate values ​​relative to the central reference plane as an optimization target, iteratively optimizing the compensation deviation parameter includes:

[0032] The distance standard deviation of the vertical distance value from each compensation sampling point to the central reference plane is determined by each compensation coordinate value as an objective function, and the compensation deviation parameter is iteratively optimized until the distance standard deviation is minimized.

[0033] In an optional embodiment of the present application, iteratively optimizing the compensation deviation parameter until the distance standard deviation is minimized includes:

[0034] Iteratively optimizing the compensation deviation parameter until the distance standard deviation is minimized;

[0035] When the variation norm of the compensation deviation parameter in two consecutive optimization iterations is lower than a preset threshold, the iteration process ends.

[0036] The present invention provides a deviation calibration method for a laser radar, wherein the laser radar is connected to a pan-tilt for driving the laser radar to rotate; the deviation calibration method comprises: controlling the laser radar to perform point cloud scanning on a measurement plane, obtaining the sampling coordinate value of each sampling point in the laser radar spherical coordinate system and obtaining the angle parameter of the pan-tilt driving the laser radar to rotate corresponding to each sampling point; performing compensation conversion on each sampling coordinate value according to the conversion relationship between the laser radar spherical coordinate system and the pan-tilt three-dimensional rectangular coordinate system, the set compensation deviation parameter and the angle parameter corresponding to each sampling point, obtaining the compensation coordinate value of the compensation sampling point corresponding to each sampling point in the pan-tilt three-dimensional rectangular coordinate system; determining the central reference plane of each compensation sampling point according to the compensation coordinate value; taking the deviation of each compensation sampling point relative to the central reference plane as the optimization target, iteratively optimizing the compensation deviation parameter until the compensation deviation parameter corresponding to the minimum deviation is obtained as the calibration deviation parameter.

[0037] In this application, the laser radar is controlled to scan the point cloud of the measurement plane, and after obtaining the sampling coordinate values ​​of each sampling point on the measurement plane, when the entire scanning system is completely error-free, each sampling coordinate value is converted to the coordinate value in the three-dimensional rectangular coordinate system of the pan-tilt head, and each position point represented by the coordinate value should also be located in the same plane; therefore, based on this, this application sets the compensation deviation parameter for compensating the deviation of the entire scanning system, and compensates the coordinate value in the process of converting the sampling coordinate value of each sampling point into the three-dimensional rectangular coordinate system of the pan-tilt head to obtain the compensation coordinate value, and then verifies the accuracy of the compensation deviation parameter based on the deviation between each compensation sampling point and the central reference plane, thereby optimizing the compensation deviation parameter, and finally determining a set of compensation deviation parameters that can make the compensation sampling points after compensation have little deviation from the central reference plane as calibration deviation parameters. When the laser radar is used to scan the scene objects in the three-dimensional space in the future, the calibration deviation parameter is used to compensate and correct the collected point cloud, so as to ensure the accuracy of the measurement of the object scene target, which is conducive to the wide application of the laser radar. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A schematic diagram of a process flow of a laser radar deviation calibration method provided in an embodiment of the present application;

[0040] Figure 2 A schematic diagram of the structure of the laser radar scanning point cloud provided for this application;

[0041] Figure 3 A schematic diagram of the distribution structure of the laser inside the laser radar provided in the embodiment of the present application;

[0042] Figure 4 A lateral schematic diagram of the relative positional relationship between the compensation sampling points and the central reference plane provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] The core of the present invention is to provide a laser radar deviation calibration method, which can improve the accuracy of laser radar measurement of target objects to a certain extent.

[0044] In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] like Figures 1 to 4 As shown, Figure 1 A schematic diagram of a process flow of a laser radar deviation calibration method provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of the laser radar scanning point cloud provided for this application; Figure 3 A schematic diagram of the distribution structure of the laser inside the laser radar provided in the embodiment of the present application; Figure 4 A lateral schematic diagram of the relative positional relationship between the compensation sampling points and the central reference plane provided in an embodiment of the present application.

[0046] In a specific embodiment of the present application, the deviation calibration method of the laser radar may include:

[0047] S1: Control the laser radar to scan the point cloud of the measurement plane, obtain the sampling coordinate value of each sampling point in the laser radar spherical coordinate system, and obtain the corresponding angle parameter of the gimbal driving the laser radar to rotate when collecting each sampling point;

[0048] S2: According to the conversion relationship between the laser radar spherical coordinate system and the gimbal three-dimensional rectangular coordinate system, the set compensation deviation parameters and the angle parameters corresponding to each sampling point, each sampling coordinate value is compensated and converted to obtain the compensation coordinate value of the compensation sampling point corresponding to each sampling point in the gimbal three-dimensional rectangular coordinate system;

[0049] S3: Determine the central reference plane of each compensation sampling point according to the compensation coordinate value;

[0050] S4: Taking the deviation of each compensation sampling point relative to the central reference plane as the optimization target, iteratively optimize the compensation deviation parameter until the compensation deviation parameter corresponding to the minimum deviation is obtained as the calibration deviation parameter.

[0051] Reference Figure 2 In order to calibrate the laser radar in this application, the wall surface or other plane is used as the measurement plane, and each position point on the measurement plane is scanned and collected. Obviously, each position point on the measurement plane must be located on the same plane. Then, in the process of the laser radar measuring each position point on the measurement plane, if the laser radar's measurement is completely free of deviation, then it is obvious that the sampling points collected by the laser radar must all be located in the same plane, which can be used as a standard for determining whether the coordinate value after compensation correction is accurate in the subsequent application.

[0052] On this basis, further reference Figure 2 and Figure 3 , the laser radar of the present application is provided with a plurality of lasers arranged along a set straight line direction. It is understandable that for each laser, it is possible to realize the acquisition and measurement of a position point on the measurement plane, and a plurality of lasers arranged in a straight line can realize the acquisition and measurement of a plurality of position points along the dimension of the set straight line direction. In practical applications, in order to expand the measurement range of each laser in the dimension of the set straight line direction, the angle between the scanning direction of each laser located in the middle position to the two end positions and the set straight line direction gradually increases, that is, the scanning direction of each laser is roughly divergent outward.

[0053] In addition, each laser can be set together on the same bearing, and the bearing is connected to the rotating shaft of the rotating motor, so that the rotating motor can drive the lasers to rotate synchronously; in actual applications, the rotating shaft of the rotating motor can be parallel to the above-mentioned set straight line direction, and the rotating motor can drive each laser to rotate about the first rotating axis, and the arrangement direction of the first rotating axis and each laser is parallel to each other. Obviously, as the rotating motor drives the rotation of each laser, each laser in the laser radar can realize collection and measurement in the second dimension.

[0054] In addition, in order to further improve the measurement range of the laser radar in three-dimensional space, the present application further sets the laser radar on a gimbal that can drive the laser radar to rotate, so that when the laser radar rotates with the driving of the gimbal, each laser can realize the collection and measurement in the third dimension. The gimbal in this embodiment drives the laser radar to rotate as a whole around the second rotation axis, and the second rotation axis should be perpendicular to the first rotation axis. Figure 2 As shown, in Figure 2In the embodiment shown, the second rotation axis is a rotation axis in the horizontal direction; therefore, the first rotation axis for the rotation of each laser should be located in a vertical plane perpendicular to the second rotation axis, and because the gimbal is used to drive the laser radar as a whole to rotate about the second rotation axis, the first rotation axis should also rotate and change in the vertical plane with the second rotation axis as the center as the laser radar rotates.

[0055] Based on the above-mentioned structure of the laser radar and the discussion of the scanning measurement method, the position information of each sampling point relative to the laser radar collected by the laser radar scanning can be represented by three different sampling coordinate values ​​of distance, polar angle and azimuth angle based on the laser radar spherical coordinate system; at the same time, when the laser radar is collecting and scanning sampling points, the gimbal will drive the laser radar to rotate to different positions. Therefore, the position information of each sampling point in space needs to be represented by the sampling coordinate value corresponding to the sampling point and the angle parameter (that is, the roll angle) that the gimbal drives the laser radar to rotate when collecting the sampling point.

[0056] In order to better characterize the position of each sampling point in space, in practical applications, a three-dimensional rectangular coordinate system of the laser radar and a three-dimensional rectangular coordinate system of the gimbal can be further established based on the laser radar spherical coordinate system of the laser radar itself.

[0057] It should be noted that the laser radar three-dimensional rectangular coordinate system in this embodiment is a three-dimensional orthogonal coordinate system corresponding to the laser radar spherical coordinate system, that is, the origin of the laser radar three-dimensional rectangular coordinate system coincides with the origin of the laser radar spherical coordinate system; the three coordinate axes of the laser radar three-dimensional rectangular coordinate system are respectively axis, axis, axis, and The axis coincides with the first rotation axis, and The axis and the second rotation axis coincide; then the coordinate value of any point in the three-dimensional space in the laser radar ball coordinate system and the coordinate values ​​in the laser radar three-dimensional rectangular coordinate system The conversion relationship is satisfied:

[0058] ;

[0059] That is to say, in the sampling coordinate value of each sampling point, the distance is the straight-line distance between the sampling point and the origin of the LiDAR spherical coordinate system, and the polar angle is the line between the sampling point and the origin of the LiDAR spherical coordinate system and the LiDAR 3D rectangular coordinate system axis; azimuth is the line between the sampling point and the origin of the LiDAR spherical coordinate system and the LiDAR 3D rectangular coordinate system The angle between the axes.

[0060] In addition, based on the above discussion, the laser radar three-dimensional rectangular coordinate system and the laser radar spherical coordinate system are both coordinate systems that are stationary relative to the laser radar, but in the process of scanning and collecting each sampling point, the laser radar will drive the laser radar to rotate as a whole. It can be seen that the laser radar three-dimensional rectangular coordinate system and the laser radar spherical coordinate system are both motion coordinate systems that rotate relative to the gimbal. Therefore, this application further creates a gimbal three-dimensional rectangular coordinate system that is stationary relative to the gimbal, and the coordinate values ​​of each sampling point can be further converted into coordinate values ​​in the three-dimensional rectangular coordinate system of the gimbal, that is, the position of each sampling point in space is represented by the coordinate value in the stationary coordinate system in the three-dimensional space.

[0061] Because the laser radar rotates around the second rotation axis passing through the origin of the laser radar's three-dimensional rectangular coordinate system during the gimbal-driven rotation of the laser radar, the origin of the laser radar's three-dimensional rectangular coordinate system and the laser radar's spherical coordinate system will not move as the laser radar rotates. The axis will not move; for this reason, in order to simplify the conversion relationship between the coordinate systems, the origin of the gimbal three-dimensional rectangular coordinate system created in this application can coincide with the origin of the laser radar three-dimensional rectangular coordinate system; on this basis, among the X-axis, Y-axis, and Z-axis of the gimbal three-dimensional rectangular coordinate system, the Y-axis can coincide with the laser radar three-dimensional rectangular coordinate system, and the X-axis is the horizontal coordinate axis, and the Z-axis is the vertical coordinate value. Based on the above-mentioned way of driving the laser radar to rotate, it can be seen that as the laser radar rotates, Axis and The axis should rotate in the XOY plane, so it can be defined as When the A-axis coincides with the X-axis, the corresponding roll angle of the laser radar, that is, the angle parameter, is 0, which means that the laser radar is at the zero position of the rotational motion.

[0062] Based on the creation of the above three coordinate systems, the present application can further set compensation deviation parameters.

[0063] As described above, in the process of scanning and measuring the point cloud, there is both a rotational movement of the laser driving the laser inside the laser radar and a rotational movement of the laser radar as a whole driven by the gimbal. Therefore, in order to more accurately and reasonably compensate for the errors introduced by the two different rotational movements in the sampling process, the present application sets internal compensation deviation parameters and axis compensation deviation parameters for the compensation deviation parameters respectively; among which, the internal compensation deviation parameters are used to compensate for the deviation introduced by driving each laser in the laser radar to rotate along the first rotation axis; and the axis compensation deviation parameters are used to compensate for the deviation introduced by the gimbal driving the laser radar as a whole to rotate about the second rotation axis.

[0064] The internal compensation deviation of the present application may include a distance compensation amount, a polar angle compensation amount and an azimuth angle compensation amount, which are respectively set to compensate for the distance, polar angle and azimuth angle.

[0065] The axis compensation deviation parameter may include a first translation compensation amount, a second translation compensation amount, a first rotation compensation angle, and a second rotation compensation angle for compensating the roll angle.

[0066] The first translation compensation amount and the second translation compensation amount are respectively the translation deviations in the Z-axis direction and the X-axis direction of the three-dimensional rectangular coordinate system of the table introduced when the table drives the laser radar to rotate;

[0067] The first rotation compensation angle and the second rotation compensation angle are respectively used to compensate for the angular deviations of the Z axis and the X axis relative to the three-dimensional rectangular coordinate system of the tabletop when the tabletop drives the laser radar to rotate.

[0068] As mentioned above, the gimbal drives the laser radar to rotate in the Y-axis and The second rotation axis coincides with the axis of rotation. Therefore, the deviation introduced by the relative motion between the laser radar and the table will not occur on the Y axis and Deviations will occur in the axis direction, and deviations will only occur in the Z-axis and X-axis directions; and, because the laser radar rotates relative to the table, this movement will not only introduce translation deviations along the Z-axis and X-axis directions, but also introduce deviations in the rotation angles relative to the Z-axis and X-axis directions. For this reason, in this embodiment, by setting the first translation compensation amount, the second translation compensation amount, the first rotation compensation angle, and the second rotation compensation angle, the accuracy and reliability of compensating for the error introduced by the rotation of the laser radar relative to the table can be guaranteed to a certain extent.

[0069] Based on the above-set internal compensation deviation parameters and axis compensation deviation parameters, the process of further compensating and converting the sampling coordinate values ​​of each sampling point into the three-dimensional rectangular coordinate system of the table may include:

[0070] S21: Compensate the sampled coordinate values ​​according to the internal compensation deviation parameters and convert them into preliminary compensated coordinate values ​​in the three-dimensional motion rectangular coordinate system of the laser radar ;in, ; For the Sampling coordinate values; They are distance compensation, polar angle compensation and azimuth angle compensation respectively;

[0071] S22: Determine the gimbal compensation matrix based on the conversion relationship between the laser radar spherical coordinate system and the gimbal coordinate system and the axis compensation deviation parameters ,in, , ; , , is the first translation compensation amount, is the second translation compensation, is the first rotation compensation angle, is the second rotation compensation angle;

[0072] S23: Convert the preliminary compensation coordinate value into the compensation coordinate value according to the pan / tilt compensation matrix and the table conversion matrix ;in, is the table transformation matrix, , For the The roll angle corresponding to each sampling point.

[0073] In this embodiment, according to the conversion relationship between the laser radar spherical coordinate system and the laser radar three-dimensional rectangular coordinate system, the sampling coordinate value of each sampling point in the laser radar spherical coordinate system is first added with the internal compensation deviation parameter, and then converted into the coordinate value in the laser radar three-dimensional rectangular coordinate system, that is, the preliminary compensation coordinate value is obtained. Obviously, the initial compensation coordinate value satisfies: ;in, For the The sampling coordinate values ​​of the sampling points; They are distance compensation, polar angle compensation and azimuth angle compensation respectively.

[0074] Furthermore, since the laser radar 3D rectangular coordinate system is a coordinate system that rotates relative to the gimbal 3D rectangular coordinate system, the gimbal compensation matrix can be further determined based on the conversion relationship between the laser radar 3D rectangular coordinate system and the gimbal 3D rectangular coordinate system and the axis compensation deviation parameters. ,in, , ; , , is the first translation compensation amount, is the second translation compensation, is the first rotation compensation angle, is the second rotation compensation angle.

[0075] Based on the gimbal compensation matrix and the conversion relationship between the three-dimensional rectangular coordinate system of the laser radar and the three-dimensional rectangular coordinate system of the gimbal, it can be further determined that the compensation coordinate values ​​converted to the three-dimensional rectangular coordinate system of the gimbal after the axis compensation deviation parameters are compensated satisfy the following: ;in It is the gimbal transformation matrix that characterizes the transformation relationship between the three-dimensional rectangular coordinate system of the laser radar and the three-dimensional rectangular coordinate system of the gimbal. Based on the way the gimbal drives the laser radar to rotate, it can be determined ;in, That is to collect The gimbal drives the lidar to rotate the corresponding roll angle.

[0076] It can be understood that if the above-mentioned internal compensation deviation parameters and axis compensation deviation parameters are set more accurately and reasonably, the compensation sampling points represented by the various compensation coordinate values ​​should basically be located in the same plane.

[0077] In order to verify each compensation deviation parameter, the central reference plane may be further determined based on the compensation coordinate values ​​corresponding to each compensation sampling point.

[0078] Reference Figure 4 The central reference plane in this application is a plane that can roughly characterize the distribution of each compensation sampling point. In practical applications, a fitting algorithm can be used to fit each compensation coordinate value to determine the fitting plane corresponding to each compensation sampling point, and the fitting plane is used as the central reference plane.

[0079] In addition, in another optional embodiment of the present application, another implementation method for determining the central reference plane is further provided, which may specifically include:

[0080] S31: Determine the centroid coordinate value corresponding to the centroid of each compensation sampling point according to each compensation coordinate value ;in, ;

[0081] S32: Construct a mathematical matrix based on the centroid coordinate value and each compensation coordinate value ;

[0082] S33: Perform singular value decomposition on the mathematical matrix to obtain a singular vector corresponding to the minimum singular value, and use a plane passing through the centroid and with the singular vector as a normal vector as a central reference plane.

[0083] In this embodiment, the central reference plane is determined by using singular value decomposition. The compensation coordinate values ​​of each compensation sampling point are averaged to obtain the centroid of each compensation sampling point. A mathematical matrix is ​​constructed based on the difference between each compensation coordinate value and the centroid coordinate value. The mathematical matrix is ​​subjected to singular value decomposition, and then:

[0084] ;in, , are all orthogonal matrices, is a diagonal matrix with singular values ​​on the diagonal. Therefore, taking the centroid as the origin, we can extract the minimum singular value The corresponding singular vectors , which is the normal vector of the central reference plane, expressed as The center reference plane can be represented by combining the center of mass coordinate value and the normal vector.

[0085] like Figure 4 As shown in FIG. 1 , the initially set compensation coordinate values ​​may not be able to accurately compensate and correct the coordinate values ​​of each sampling point, so the determined compensation sampling points fluctuate up and down on the central reference plane. Obviously, when the distance of each compensation sampling point from the central reference plane is smaller, it means that each compensation sampling point is closer to being located in the same plane. Ideally, each compensation sampling point should be located in the same plane. Therefore, in this embodiment, the size of the distance standard deviation of the vertical distance between each compensation sampling point and the central reference plane is used as a standard to measure whether the compensation deviation parameter is accurate, and the compensation deviation parameter is optimized and adjusted accordingly; Figure 4 As shown, when the compensation deviation parameter can be well adjusted for each sampling point, the distance of the finally determined compensation sampling point from the central reference plane is smaller.

[0086] Therefore, in this embodiment, the distance standard deviation of the vertical distance value from each compensation sampling point to the central reference plane determined by each compensation coordinate value can be used as the objective function, and the compensation deviation parameter can be iteratively optimized until the distance standard deviation is minimized.

[0087] In practical applications, after determining the central reference plane, we can further determine the vertical distance of each compensation sampling point relative to the central reference plane, that is, the distance from the compensation sampling point to the central plane. Obviously, this vertical distance can also characterize the deviation of the compensation sampling point relative to the central reference plane.

[0088] In the process of determining the vertical distance between each compensation sampling point and the central reference plane, the vertical distance can be obtained by multiplying the compensation coordinate value of each compensation sampling point by the normal vector of the central reference plane. ; Further calculate the average distance of the vertical distance corresponding to each compensation sampling point ; Based on the average distance, the standard deviation of the distance can be further determined as .

[0089] The distance standard deviation determined in the above implementation is the overall standard deviation of all compensation sampling points.

[0090] However, the present application further considers that the number of sampling points collected by the laser radar may be relatively large, and the number of compensation sampling points will be correspondingly large, and the amount of calculation to determine the overall standard deviation may be relatively large. For this reason, in another optional implementation of the present application, in actual applications, a number of points can also be randomly selected from each compensation sampling point as sample sampling points; thus, in the process of determining the distance standard deviation, only the average vertical distance between each sample sampling point and the central reference plane can be calculated, and then the sample standard deviation can be determined based on the coordinate value of each sample sampling point and the average vertical distance; that is, the sample standard deviation is: ;in, is the total number of sample sampling points selected; for Obviously, whether the overall standard deviation of vertical distance or the sample standard deviation is used as the distance standard deviation to judge whether the compensation deviation parameter is accurate does not affect the implementation of the technical solution of the present application.

[0091] In the process of iteratively optimizing the compensation deviation parameters with the above distance standard deviation as the objective function, the Levenberg-Marquardt algorithm can be used to iteratively optimize the compensation deviation parameters, or other optimization algorithms that can automatically calculate the gradient or do not require explicit gradient calculation can be selected until the distance standard deviation is relatively small; in addition, when the norm of the change in the compensation deviation parameter in two consecutive optimization iterations is lower than the preset threshold, the iterative process ends. For example, the preset threshold can be .

[0092] In summary, in this application, the laser radar is controlled to scan the point cloud of the measurement plane, and after obtaining the sampling coordinate values ​​of each sampling point on the measurement plane, when the entire scanning system is completely error-free, each sampling coordinate value is converted to the coordinate value in the three-dimensional rectangular coordinate system of the pan-tilt head, and each position point represented by the coordinate value should also be located in the same plane; therefore, based on this, this application sets the compensation deviation parameter for compensating the deviation of the entire scanning system, and compensates the coordinate value in the process of converting the sampling coordinate value of each sampling point into the three-dimensional rectangular coordinate system of the pan-tilt head to obtain the compensation coordinate value, and then verifies the accuracy of the compensation deviation parameter based on the deviation between each compensation sampling point and the central reference plane, thereby optimizing the compensation deviation parameter, and finally determining a set of compensation deviation parameters that can make the compensation sampling points basically have a small deviation from the central reference plane as the calibration deviation parameter. When the laser radar is used to scan the scene objects in the three-dimensional space in the future, the calibration deviation parameter is used to compensate and correct the collected point cloud, so as to ensure the accuracy of the measurement of the object scene target, which is conducive to the wide application of the laser radar.

[0093] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "includes", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment comprising a series of elements are inherent to the elements. In the absence of more restrictions, the elements limited by the sentence "comprising one..." do not exclude the presence of other identical elements in the process, method, article or equipment comprising the elements. In addition, the above-mentioned technical solution provided in the embodiment of the present application is consistent with the corresponding technical solution in the prior art in principle, and the part is not described in detail, so as to avoid too much redundancy.

[0094] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core ideas of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the present invention.

Claims

1. A laser radar deviation calibration method, characterized in that: The laser radar is connected to a pan / tilt for driving the laser radar to rotate; the deviation calibration method comprises: Control the laser radar to perform point cloud scanning on the measurement plane, obtain the sampling coordinate value of each sampling point in the laser radar spherical coordinate system, and obtain the angle parameter of the pan-tilt drive the laser radar to rotate corresponding to each sampling point; According to the conversion relationship between the laser radar spherical coordinate system and the gimbal three-dimensional rectangular coordinate system, the set compensation deviation parameter and the angle parameter corresponding to each sampling point, each sampling coordinate value is compensated and converted to obtain the compensation coordinate value of the compensation sampling point corresponding to each sampling point in the gimbal three-dimensional rectangular coordinate system; Determine the central reference plane of each of the compensation sampling points according to the compensation coordinate value; The deviation of each compensation sampling point relative to the central reference plane is taken as an optimization target, and the compensation deviation parameter is iteratively optimized until the compensation deviation parameter corresponding to the minimum deviation is obtained as a calibration deviation parameter.

2. The deviation calibration method of the rotating laser radar according to claim 1, characterized in that: The laser radar includes a plurality of lasers arranged in sequence along a set straight line direction; Controlling the laser radar to perform point cloud scanning on the measurement plane includes: The laser in the laser radar is controlled to rotate around the first rotation axis, and the laser radar is driven to rotate around the second rotation axis through the gimbal, and each laser is synchronously controlled to perform point cloud scanning on the position points on the measurement plane; wherein the first rotation axis and the set straight line direction are parallel to each other, and the first rotation axis and the second rotation axis are perpendicular to each other.

3. The deviation calibration method of the rotating laser radar as claimed in claim 2, characterized in that: From the middle position to the two end positions of each laser, the angle between the scanning direction of each laser and the set straight line direction gradually decreases; The second rotation axis is a horizontal rotation axis.

4. The deviation calibration method of the rotating laser radar according to claim 2 or 3, characterized in that: The sampling coordinate values ​​include distance, polar angle and azimuth; the angle parameters include roll angle; Wherein, the distance is the straight-line distance between the sampling point and the origin of the laser radar spherical coordinate system; The polar angle and the azimuth are the angles of the line between the sampling point and the origin of the laser radar spherical coordinate system and the laser radar 3D rectangular coordinate system. Axis and The angle between the axes; wherein the origin of the laser radar three-dimensional rectangular coordinate system coincides with the origin of the laser radar spherical coordinate system; The axis coincides with the first rotation axis, the The axis coincides with the second rotation axis; The roll angle is the angle at which the gimbal controls the laser radar to rotate around the second rotation axis when scanning the sampling point.

5. The deviation calibration method of the rotating laser radar as claimed in claim 4, characterized in that: The laser radar spherical coordinate system and the laser radar three-dimensional rectangular coordinate system are two relatively stationary coordinate systems, and both are coordinate systems that move with respect to the table three-dimensional rectangular coordinate system about the second rotation axis; The X-axis direction and the Y-axis direction of the three-dimensional rectangular coordinate system of the table are both horizontal directions, and the Z-axis direction is the vertical direction; and the Y-axis of the three-dimensional rectangular coordinate system of the table and the three-dimensional rectangular coordinate system of the laser radar are Axes coincide.

6. The deviation calibration method of the rotating laser radar as claimed in claim 5, characterized in that: The compensation deviation parameters set include internal compensation deviation parameters and axis compensation deviation parameters; The internal compensation deviation parameters include a distance compensation amount, a polar angle compensation amount and an azimuth angle compensation amount respectively set for compensating the distance, the polar angle and the azimuth angle; The axis compensation deviation parameters include a first translation compensation amount, a second translation compensation amount, a first rotation compensation angle, and a second rotation compensation angle for compensating the roll angle; Among them, the first translation compensation amount and the second translation compensation amount are respectively used to compensate for the translation deviations in the Z-axis direction and the X-axis direction of the table top three-dimensional rectangular coordinate system introduced when the table top drives the laser radar to rotate; the first rotation compensation angle and the second rotation compensation angle are respectively used to compensate for the angular deviations relative to the Z-axis and the X-axis of the table top three-dimensional rectangular coordinate system introduced when the table top drives the laser radar to rotate.

7. The deviation calibration method of the rotating laser radar as claimed in claim 6, characterized in that: According to the conversion relationship between the laser radar spherical coordinate system and the gimbal coordinate system and the set compensation deviation parameter, each of the sampled coordinate values ​​is compensated and converted into a compensated coordinate value in the gimbal coordinate system, including: The sampled coordinate values ​​are compensated according to the internal compensation deviation parameters and then converted into preliminary compensated coordinate values ​​in the three-dimensional motion rectangular coordinate system of the laser radar. ;in, ; For the The sampling coordinate values ​​of the sampling points; are the distance compensation amount, the polar angle compensation amount and the azimuth angle compensation amount respectively; Determine the gimbal compensation matrix according to the conversion relationship between the laser radar three-dimensional rectangular coordinate system and the gimbal three-dimensional rectangular coordinate system and the axis compensation deviation parameter ,in, , ; , , is the first translation compensation amount, is the second translation compensation amount, is the first rotation compensation angle, is the second rotation compensation angle; The preliminary compensation coordinate value is converted into the compensation coordinate value according to the pan-tilt compensation matrix and the table conversion matrix ;in, is the table transformation matrix, , For the The roll angle corresponding to each of the sampling points.

8. The deviation calibration method of the rotating laser radar as claimed in claim 7, characterized in that: Determining a central reference plane corresponding to the compensation value of each sampling point according to the compensation coordinate value includes: A fitting plane corresponding to each of the compensation sampling points is determined according to each of the compensation coordinate values, and the fitting plane is used as the central reference plane.

9. The deviation calibration method of the rotating laser radar as claimed in claim 8, characterized in that: Taking the deviation value of each of the compensation coordinate values ​​relative to the central reference plane as an optimization target, iteratively optimizing the compensation deviation parameter includes: The distance standard deviation of the vertical distance value from each compensation sampling point to the central reference plane is determined by each compensation coordinate value as an objective function, and the compensation deviation parameter is iteratively optimized until the distance standard deviation is minimized.

10. The deviation calibration method of the rotating laser radar according to claim 8, characterized in that: Iteratively optimizing the compensation deviation parameter until the distance standard deviation is minimized includes: The compensation deviation parameter is iteratively optimized, and when the norm of the change amount of the compensation deviation parameter in two consecutive optimization iterations is lower than a preset threshold, the iterative process ends.

Images