Method and device for flight calibration of inter-channel angle error of multi-channel laser sensor

Abstract

The application discloses a multi-channel laser sensor inter-channel angle error flight calibration method, device and equipment and a computer readable storage medium, and belongs to the technical field of unmanned aerial vehicle laser remote sensing. The method comprises the following steps: calibrating and splitting obtained point cloud data files to obtain channel point cloud data files; determining feature surface patches and reference surface patches based on channel point cloud data in the channel point cloud data files, wherein the reference surface patches are feature surface patches corresponding to middle channels; calculating channel angle errors of each channel according to distances from feature points on the feature surface patches to reference points on the reference surface patches, and determining angle error calibration results based on the channel angle errors and theoretical errors. In this way, angle error calibration is performed on each channel of the multi-channel laser sensor, and the precision of the multi-channel laser sensor is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle laser remote sensing, and particularly relates to a multi-channel laser sensor inter-channel angle error flight calibration method, device and equipment and a computer readable storage medium. BACKGROUND

[0002] Unmanned aerial vehicle airborne laser remote sensing technology is a result of the increasing maturity of unmanned aerial vehicles and laser sensor technology, and is the most commonly used topographic survey method in the current surveying and mapping field. In recent years, with the development of the automatic driving industry, the types of laser sensors have increased, and the performance has improved while the cost has gradually decreased, and many products using vehicle-level multi-channel laser sensors integrated as airborne surveying and mapping laser loads have appeared in the industry.

[0003] Compared with surveying and mapping level single-channel laser sensors, multi-channel laser sensors have many shortcomings, such as large spot size, unclear details, difficulty in accurate calibration between multi-channels according to surveying and mapping accuracy, and large angle and distance measurement accuracy deviation, and therefore can only be used for small scale topographic map measurement with low accuracy requirements.

[0004] Among the many errors of multi-channel lasers, the horizontal and vertical angle errors between channels have the greatest impact on the thickness and accuracy of the results. In existing products, most directly use the calibration results provided by laser sensor manufacturers without surveying and mapping level calibration. The manufacturers produce laser sensors to adapt to the short distance requirements of automatic driving, and usually perform short distance calibration of multi-channel lasers within about 30m in the laboratory. The calibration results cannot be stably used for long distance surveying and mapping. Some surveying and mapping instrument integrators build ground control fields by themselves, and use high-precision prior point clouds of buildings for equipment calibration, but the scanning lines cannot be concentrated at a long distance, and the final calibration results are still not ideal, and the operation is complex, and cannot achieve long distance accurate and efficient calibration effect.

[0005] The above content is only used to assist in understanding the technical solutions of the present application, and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0006] The main purpose of the present application is to provide a multi-channel laser sensor inter-channel angle error flight calibration method, device, equipment and computer readable storage medium, which aims to improve the accuracy of multi-channel laser sensors.

[0007] To achieve the above purpose, the present application provides a multi-channel laser sensor inter-channel angle error flight calibration method, comprising the following steps:

[0008] A multi-channel laser sensor inter-channel angle error flight calibration method comprises:

[0009] Calibrate and split the obtained point cloud data file to obtain a channel point cloud data file.

[0010] determining a feature patch and a reference patch based on the channel point cloud data in the channel point cloud data file, wherein the reference patch is the feature patch corresponding to the intermediate channel;

[0011] calculating a channel angle error of each channel according to the distance from the feature point on the feature patch to the reference point on the reference patch, and determining an angle error calibration result based on the channel angle error and a theoretical error.

[0012] Optionally, the calibration and splitting of the obtained point cloud data file to obtain the channel point cloud data file comprises:

[0013] adjusting and calibrating the obtained original point cloud data to obtain the overall installation error of the IMU and the laser sensor body;

[0014] calibrating the original point cloud data based on the overall installation error to obtain calibrated point cloud data;

[0015] splitting the calibrated point cloud data according to the flight lines to obtain flight line point cloud data files, the number of the flight line point cloud data files being consistent with the number of the flight lines;

[0016] splitting any one of the flight line point cloud data files according to the scanning lines to obtain channel point cloud data files, the number of the channel point cloud data files being consistent with the number of the channels.

[0017] Optionally, the determination of the feature patch and the reference patch based on the channel point cloud data in the channel point cloud data file comprises:

[0018] dividing the channel point cloud data into blocks to obtain channel point cloud data blocks;

[0019] calculating the normal vector and the roughness of each feature point in the channel point cloud data block;

[0020] determining a seed point based on the roughness, and performing plane point cloud growth based on the seed point to obtain the feature patch;

[0021] selecting a feature patch of an intermediate channel from the obtained feature patches as the reference patch.

[0022] Optionally, the calculation of the channel angle error of each channel according to the distance from the feature point on the feature patch to the reference point on the reference patch, and the determination of the angle error calibration result based on the channel angle error and the theoretical error comprise:

[0023] taking the sum of squares of the distances from the feature point on the feature patch to the reference point on the reference patch as a constraint condition, and calculating the channel angle error of each channel;

[0024] Superimpose the channel angle error of each channel with the theoretical error respectively, and determine the superimposed result as the angle error calibration result of the corresponding channel.

[0025] Optionally, the calculating the channel angle error of each channel based on the square sum of the distances from the feature points on the feature patch to the reference points on the reference patch as a constraint condition comprises:

[0026] calculating the feature coordinates of the feature points on the feature patch based on the original coordinates of the laser foot points in the original coordinate system of the laser sensor, the rotation matrix, and the coordinates of the feature points in the scene coordinate system;

[0027] determining a first distance equation of the feature points to the reference plane based on the feature coordinates and the normal vector of the feature points;

[0028] linearizing the first distance equation using a Taylor expansion to obtain a second distance equation;

[0029] obtaining a coefficient matrix of unknowns based on the second distance equation, and determining an error equation based on the coefficient matrix of unknowns and the distances of the laser foot points to the reference plane;

[0030] determining the angle error of the corresponding channel based on the error equation and the least square method principle.

[0031] Optionally, before the calibrating and splitting the obtained point cloud data file to obtain the channel point cloud data file, the method further comprises:

[0032] controlling the unmanned aerial vehicle to fly along a preset flight path on a pre-arranged site to obtain the point cloud data file, wherein the unmanned aerial vehicle is mounted with the multi-channel laser sensor to be calibrated and an inertial navigation device.

[0033] Optionally, after the calculating the channel angle error of each channel based on the distances from the feature points on the feature patch to the reference points on the reference patch, and determining the angle error calibration result based on the channel angle error and the theoretical error, the method further comprises:

[0034] calibrating each channel based on the angle error calibration result to complete the calibration of the inter-channel angle error of the multi-channel laser sensor.

[0035] To achieve the above-mentioned purposes, the application further provides a flight calibration device for inter-channel angle error of a multi-channel laser sensor, comprising:

[0036] a file obtaining module configured to calibrate and split an obtained point cloud data file to obtain a channel point cloud data file;

[0037] The face sheet determination module is configured to determine feature face sheets and reference face sheets based on the channel point cloud data in the channel point cloud data file, wherein the reference face sheets are feature face sheets corresponding to the intermediate channels.

[0038] The angle error calculation module is configured to calculate channel angle errors of the respective channels according to distances from feature points on the feature face sheets to reference points on the reference face sheets, and determine an angle error calibration result based on the channel angle errors and a theoretical error.

[0039] To achieve the above object, the present application further provides a device for flight calibration of inter-channel angle errors of a multi-channel laser sensor, which comprises a memory, a processor, and a flight calibration program of inter-channel angle errors of a multi-channel laser sensor stored in the memory and executable on the processor, and the flight calibration program of inter-channel angle errors of a multi-channel laser sensor, when executed by the processor, implements the steps of the flight calibration method of inter-channel angle errors of a multi-channel laser sensor.

[0040] To achieve the above object, the present application further provides a computer readable storage medium, which stores a flight calibration program of inter-channel angle errors of a multi-channel laser sensor, and the flight calibration program of inter-channel angle errors of a multi-channel laser sensor, when executed by a processor, implements the steps of the flight calibration method of inter-channel angle errors of a multi-channel laser sensor.

[0041] Compared with the prior art, the present application discloses a flight calibration method, device, equipment and computer readable storage medium of inter-channel angle errors of a multi-channel laser sensor, which comprises: calibrating and splitting an obtained point cloud data file to obtain a channel point cloud data file; determining feature face sheets and reference face sheets based on channel point cloud data in the channel point cloud data file, wherein the reference face sheets are feature face sheets corresponding to the intermediate channels; calculating channel angle errors of the respective channels according to distances from feature points on the feature face sheets to reference points on the reference face sheets, and determining an angle error calibration result based on the channel angle errors and a theoretical error. Thus, the angle error calibration is performed on each channel of the multi-channel laser sensor, and the precision of the multi-channel laser sensor is improved. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 Flowchart of an embodiment of the flight calibration method of inter-channel angle errors of a multi-channel laser sensor of the present application;

[0043] Figure 2 Composition diagram of a UAV mounting device related to an embodiment of the present application;

[0044] Figure 3 Internal structure diagram of a multi-channel laser sensor related to an embodiment of the present application;

[0045] Figure 4 The figure shows the angle between the channels of the multi-channel laser sensor according to an embodiment of the present application;

[0046] Figure 5 The figure shows the vertical angle and the horizontal angle of the multi-channel laser sensor according to an embodiment of the present application;

[0047] Figure 6 The figure shows the marker according to an embodiment of the present application;

[0048] Figure 7 The figure shows the detailed flow of the flight calibration method for the angle error between the channels of the multi-channel laser sensor according to an embodiment of the present application;

[0049] Figure 8 The figure shows the scene of the flight calibration method for the angle error between the channels of the multi-channel laser sensor according to an embodiment of the present application;

[0050] Figure 9 The figure shows the structure of the flight calibration device for the angle error between the channels of the multi-channel laser sensor according to an embodiment of the present application;

[0051] Figure 10 The figure shows the internal structure of the device according to an embodiment of the present application.

[0052] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0053] The embodiments of the present application will be described in detail below, and the examples of the embodiments are shown in the accompanying drawings, in which the same or similar notations represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application. All other embodiments obtained by those skilled in the art based on the embodiments in the present application without creative labor are within the scope of protection of the present application.

[0054] The first embodiment of the present application proposes a flight calibration method for the angle error between the channels of a multi-channel laser sensor, which is described with reference to Figure 1 , Figure 1 The figure shows the flow of the flight calibration method for the angle error between the channels of the multi-channel laser sensor according to an embodiment of the present application, which includes: Figure 1

[0055] In step S101, the obtained point cloud data file is calibrated and split to obtain a channel point cloud data file;

[0056] ​The point cloud data file comprises point cloud data obtained during flight of the unmanned aerial vehicle. The point cloud data file is obtained by controlling the unmanned aerial vehicle to fly along a preset flight path on a pre-arranged site. Referring to Figure 2 , Figure 2 An embodiment of the present application relates to a diagram of a mounting device of an unmanned aerial vehicle. In this embodiment, a multi-channel laser sensor 11, an inertial navigation device 12, an onboard computer 13, and a global navigation satellite system (GNSS) device 14 are mounted on an unmanned aerial vehicle 100. The onboard computer 13 is connected to the GNSS device 14, the inertial navigation device 12, and the multi-channel laser sensor 11, and the connection can be electrical or wireless. The laser sensor 11 and the inertial navigation device are arranged on a gimbal of the unmanned aerial vehicle. The inertial navigation device integrates an inertial measurement unit (IMU), a barometer, a global positioning system (GPS), a magnetic compass, and other sensors.

[0057] The unmanned aerial vehicle 100 flies along a pre-designed 3x3 vertical cross flight path. The GNSS device 14 records the position of the unmanned aerial vehicle platform at a rate of 20 Hz, the inertial navigation device 12 records the attitude change of the multi-channel laser sensor 11 at a rate of 300 Hz, and the onboard computer 13 is used to achieve time synchronization of the GNSS device 14, the inertial navigation device 12, and the multi-channel laser sensor 11, and records all data. The time synchronization accuracy is less than 1 ms. Referring to Figure 3 , Figure 3 An embodiment of the present application relates to a diagram of the internal structure of a multi-channel laser sensor. The multi-channel laser sensor 11 includes an internal rotation shaft 111, a plurality of laser emitters 112, and a lens 113. Each laser emitter 112 is arranged and installed at a preset angle. The laser emitters 112 and the lens 113 are oppositely arranged on both sides of the internal rotation shaft 111. Further referring to Figure 4 The plurality of laser emitters 112 and the lens 113 form a scanning line 114 around the rotation shaft, achieving multi-channel scanning. The scanning line 114 formed by the laser emitted by the plurality of laser emitters 112 after passing through the lens 113 corresponds to the laser emitters 112.

[0058] The theoretical installation angle between the laser emitters and the multi-channel laser sensor is pre-set, but due to the error of electronic welding, there are errors in the vertical and horizontal directions as shown in Figure 5 The manufacturer will provide a theoretical error to ensure normal use in short-distance scanning. This embodiment further corrects the angle error in a long-distance scene to improve the accuracy of the long-distance scene.

[0059] In order to control the point cloud data algorithm through the markers on the site, a plurality of markers are arranged at a long distance on the site, and the markers have 5 upward and different surfaces, such as Figure 6 In this embodiment, the bottom side length of the marker is 2m, and the side inclination angle is 45 degrees. In other embodiments, markers of different shapes and sizes can be set based on actual needs.

[0060] When the unmanned aerial vehicle device and the long-distance site are ready, the multi-channel laser sensor 11, the inertial navigation device 12, the on-board computer 13 and the GNSS device 14 are controlled to run, and the unmanned aerial vehicle 100 is controlled to take off, so that the unmanned aerial vehicle 100 flies according to the pre-designed 3x3 vertical cross flight path. The point cloud data obtained during flight is recorded by the on-board computer 13 and saved as a point cloud data file for further processing of the point cloud data file.

[0061] In this embodiment, step S101 includes:

[0062] The obtained original point cloud data is adjusted and calibrated to obtain the overall installation error of the IMU and the laser sensor body;

[0063] The original point cloud data is calibrated based on the overall installation error to obtain calibrated point cloud data;

[0064] The calibrated point cloud data is split according to the flight path to obtain flight path point cloud data files, and the number of flight path point cloud data files is consistent with the number of flight paths;

[0065] Any one flight path point cloud data file is split according to the scanning line to obtain channel point cloud data files, and the number of channel point cloud data files is consistent with the number of channels.

[0066] In this example, the adjustment and calibration are based on the automatic geometric step method. This method takes the installation error between the IMU and the laser sensor body in the roll, pitch and heading directions as unknowns, and through the coincidence characteristics of the ground objects between different flight strips, the overall installation error of the IMU and the laser sensor body is obtained through the existing algorithm according to the geometric constraint automatic step adjustment and calibration.

[0067] After obtaining the overall installation error, the original point cloud data is calibrated based on the overall installation error to obtain calibrated point cloud data. This embodiment sets a 3x3 (three horizontal and three vertical) vertical intersection flight, and therefore also needs to split the calibrated point cloud data based on the flight to obtain six flight point cloud data files, which correspond to the six flights respectively. Any one flight point cloud data file is split according to the scanning line to obtain a channel point cloud data file, and the number of channel point cloud data files is consistent with the number of channels. The flight point cloud data file includes the channel number of the laser emitter, and the channel numbers consistent with each other belong to the same category. The point cloud data in the flight point cloud data file is classified based on the channel number, and the point cloud data of the same channel number is classified into the same category. After classification, each category is saved as a channel point cloud data file. The number of channel point cloud data files is consistent with the number of channels, for example, n channels correspond to n channel point cloud data files.

[0068] In step S102, a feature surface patch and a reference surface patch are determined based on the channel point cloud data in the channel point cloud data file, wherein the reference surface patch is a feature surface patch corresponding to the middle channel.

[0069] Specifically, the channel point cloud data is divided into blocks to obtain channel point cloud data blocks.

[0070] The normal vector and roughness of each feature point in the channel point cloud data block are calculated.

[0071] A seed point is determined based on the roughness, and a plane point cloud is grown based on the seed point to obtain a feature surface patch.

[0072] An arbitrary feature surface patch of the middle channel is selected from the obtained feature surface patches as a reference surface patch.

[0073] Due to the influence of the available memory capacity in the running environment, a large amount of point cloud data cannot be directly read into the running environment, and the channel point cloud data needs to be processed by block. Specifically, the channel point cloud data can be divided into blocks according to the available memory capacity of the running environment, the capacity of the channel point cloud data and the block size to obtain multiple channel point cloud data blocks.

[0074] The embodiment estimates the normal vector of each feature point by principal component analysis of the covariance matrix of the coordinates of adjacent points, and the normal vector is specifically a surface normal vector. The embodiment marks points in a channel point cloud data block as feature points. Specifically, an octree index of the current point cloud data block is constructed, and according to the structure of the constructed octree index, a fixed value is selected as a search radius according to a preset rule based on point density and flight path terrain, and neighborhood feature points are selected according to the search radius; the normal vector and roughness of all feature points in the current point cloud data block are calculated according to all feature points in the current point cloud data block and the neighborhood feature points corresponding to all feature points. The normal vector is calculated according to the distribution of neighborhood feature points. If the neighborhood feature points around the feature point basically form a plane, the normal vector of the feature point is a vector perpendicular to the plane, and the roughness of the plane is small, and the plane can be considered to be smooth. By constructing the octree, it is not necessary to traverse all point cloud data, and the search for neighborhood feature points can be accelerated.

[0075] Given a matrix P containing n feature point coordinates, P = {p1, p2, p3,..., pn}, (p represents three-dimensional coordinates), C(P) is a 3x3 covariance matrix of n feature points. The principal components of C(P) are eigenvectors, forming an orthogonal basis. Let λ1, λ2, λ3 represent eigenvalues, and the eigenvalues λ1, λ2, λ3 correspond to eigenvectors in the direction of the variance. Assuming that the eigenvalues are sorted in size as λ1≥ λ2≥ λ3, the third eigenvector is the surface normal vector of the surface The square root of the third eigenvalue λ3 corresponds to the standard deviation of the current feature point (i.e., the feature point whose estimated normal vector is indicated) from the surface, and therefore the third eigenvalue λ3 can be understood as a measure of roughness σ p where

[0076] The roughness of each feature point in the channel point cloud data block is sorted, and the feature point with the smallest roughness is determined as a seed point. A typical region growing algorithm is used to grow a plane point cloud based on the seed point. After completing the plane point cloud growth of a seed point, the point with the smallest roughness is selected from the remaining feature points in the channel point cloud data block as the next seed point to continue the plane point cloud growth. After each subsequent plane point cloud growth is completed, it is determined whether the number of remaining feature points is greater than or equal to a pre-set number threshold. If the number of remaining feature points is greater than or equal to the number threshold, the seed point is continued to be determined from the remaining feature points, and the plane growth operation is performed again. The process is repeated until the number of remaining feature points is less than the number threshold, or the roughness of the remaining feature points is less than a roughness threshold, and the plane region growing operation cannot be continued. The feature patches are obtained, and the point cloud data of the extracted feature patches are stored in containers respectively. Based on the above operation, the feature patches of each channel of each flight path can be obtained. A feature patch of an intermediate channel is selected as a reference patch from the obtained feature patches. For example, there are channels 1-13, and the feature patch corresponding to channel 7 can be selected as the reference plane.

[0077] In step S103, the channel angle error of each channel is calculated according to the distance from the feature point on the feature patch to the reference point on the reference patch, and the angle error calibration result is determined based on the channel angle error and the theoretical error.

[0078] Reference Figure 7 , Figure 7 The step S103 includes:

[0079] In step S1031, the channel angle error of each channel is calculated with the sum of squares of distances from the feature point on the feature patch to the reference point on the reference patch as a constraint condition. In step S1031, the feature coordinates of the feature point on the feature patch are calculated based on the original coordinates of the laser foot point in the original coordinate system of the laser sensor, the rotation matrix, and the coordinates of the feature point in the scene coordinate system. The first distance equation of the feature point to the reference plane is determined based on the feature coordinates and the normal vector of the feature point. The second distance equation is obtained by linearizing the first distance equation using the Taylor expansion. The unknown coefficient matrix is obtained based on the second distance equation, and the error equation is determined based on the unknown coefficient matrix and the distance from the laser foot point to the reference plane. The angle error of the corresponding channel is determined based on the error equation and the least squares principle.

[0080] In step S1032, the channel angle error of each channel is superimposed with the theoretical error, and the superimposed result is determined as the angle error calibration result of the corresponding channel.

[0081] The route point cloud data file can be split by scan line to obtain channel point cloud data files consistent with the number of channels. Each channel point cloud data file is a single channel and the ground point cloud data after the IMU and the overall laser sensor calibration of the integrated navigation system. Due to the slight error between different channel laser sensors, the point cloud data of multiple channels cannot completely coincide on the same point on the marker surface, and the point cloud data obtained by different laser emitters has a slight error.

[0082] Specifically, the feature patches on the surfaces of different channel point cloud sub-targets are made to coincide. In this embodiment, a least square algorithm is used to calibrate the installation angle. The feature patch of a middle channel is taken as a reference plane. If the installation angle of the channel laser emitter relative to the IMU is accurate, the distance from the feature patch of the same marker to the reference plane should be 0 in theory. Therefore, the sum of the squares of the distances from the feature points on the same feature patch to the reference plane is taken as a constraint condition to calculate the angle error between different channels.

[0083] The coordinates of the feature point P are represented as p(x, y, z). The feature coordinate determination formula of the feature point P is:

[0084]

[0085] where (x0, y0, z0) is the coordinate of the laser foot point in the original coordinate system of the laser sensor, which can be calculated by the original laser sensor observation value (point cloud data) corresponding to each feature point of the feature patch. R C R is a rotation matrix consistent with the rotation angle of the POS coordinate system. M ΔR is the overall installation error of the IMU and the laser sensor body. M is the installation angle error rotation matrix of the IMU and the channel, that is, the unknown matrix to be solved. Δx, Δy, Δz is the installation eccentricity of the laser sensor and the IMU. N R is a POS attitude selection matrix. M R is a rotation matrix for converting the NED coordinate system into the left-hand coordinate system. POS P is the coordinate of the feature point in the POS coordinate system.

[0086] Thus, the first distance equation d of the feature point P to the reference plane P ref is represented as:

[0087]

[0088] The angle error Δα, Δβ, Δγ is taken as an unknown number, and the Taylor expansion is used to linearize the above formula, and the second distance equation is obtained by ignoring the second order term and the high order term:

[0089]

[0090] Where d is the distance from the feature point to the reference surface. Let d0 be the partial derivatives of the second distance equation with respect to the unknowns Δα, Δβ, and Δγ. It should be noted that the Taylor expansion is an expansion from zero to higher orders, where d0 is the zero-order term. Ignoring second-order terms and subsequent derivatives, d0 can be considered the value of d when the laser foot point coordinates are (x0, y0, z0).

[0091] The ideal distance from the laser footpoint to the reference plane is zero, but in reality it is not zero. Let the distance residual be V, then the error equation can be obtained:

[0092] V = BX + L

[0093] B is the unknown coefficient matrix, obtained by Taylor series expansion; L is the distance matrix from the laser footpoint to the reference plane; X is the angular error [Δα, Δβ, Δγ] between the IMU and the channel to be determined. T When V T The optimal solution can be obtained when PV is minimized, where V T Let P be the transpose of matrix V, where V is the distance residual and P is the weight matrix. The angle error parameter can be obtained using the least squares principle and the following formula:

[0094] X = -(B T PB) -1 B T PL

[0095] Where X represents the angular error to be determined [Δα, Δβ, Δγ] T B is the unknown coefficient matrix, obtained by Taylor expansion; L is the distance matrix from the laser foot point to the reference plane; and P is the weight matrix.

[0096] After obtaining the angle error for each channel, the channel angle error is superimposed on the theoretical error, and the superposition result is determined as the angle error calibration result for the corresponding channel. (Reference) Figure 8 The process involves obtaining point cloud data for each channel of the laser sensor, and then calculating the channel angle errors 1-n based on the point cloud data 1-n. Theoretical errors are added to the channel angle errors 1-n to obtain the angle error calibration results for each channel. Based on these calibration results, the placement angles of each channel are calibrated, thus completing the calibration of the inter-channel angle errors of the multi-channel laser sensor.

[0097] The embodiment calibrates and splits the obtained point cloud data file based on the above scheme to obtain a channel point cloud data file; determines a feature face sheet and a reference face sheet based on channel point cloud data in the channel point cloud data file, wherein the reference face sheet is a feature face sheet corresponding to an intermediate channel; calculates a channel angle error of each channel according to a distance from a feature point on the feature face sheet to a reference point on the reference face sheet, and determines an angle error calibration result based on the channel angle error and a theoretical error. In this way, the angle error calibration of each channel of the multi-channel laser sensor is performed, and the precision of the multi-channel laser sensor is improved.

[0098] In addition, the application further provides a multi-channel laser sensor inter-channel angle error flight calibration device, which refers to Figure 9 , and the multi-channel laser sensor inter-channel angle error flight calibration device comprises:

[0099] A file obtaining module 10 is configured to calibrate and split the obtained point cloud data file to obtain a channel point cloud data file.

[0100] A face sheet determining module 20 is configured to determine a feature face sheet and a reference face sheet based on channel point cloud data in the channel point cloud data file, wherein the reference face sheet is a feature face sheet corresponding to an intermediate channel.

[0101] An angle error calculating module 30 is configured to calculate a channel angle error of each channel according to a distance from a feature point on the feature face sheet to a reference point on the reference face sheet, and determine an angle error calibration result based on the channel angle error and a theoretical error.

[0102] The file obtaining module 10 comprises:

[0103] A calibration and adjustment unit is configured to calibrate and adjust the obtained original point cloud data to obtain an overall installation error of an IMU and a laser sensor body.

[0104] An original point cloud data calibration unit is configured to calibrate the original point cloud data based on the overall installation error to obtain calibrated point cloud data.

[0105] A first splitting unit is configured to split the calibrated point cloud data according to routes to obtain route point cloud data files, and the number of the route point cloud data files is consistent with the number of routes.

[0106] A second splitting unit is configured to split any one of the route point cloud data files according to scanning lines to obtain channel point cloud data files, and the number of the channel point cloud data files is consistent with the number of channels.

[0107] Further, the face sheet determining module 20 comprises:

[0108] A block unit is configured to block the channel point cloud data to obtain channel point cloud data blocks;

[0109] A calculation unit is configured to calculate a normal vector and a roughness of each feature point in the channel point cloud data block;

[0110] A feature patch obtaining unit is configured to determine a seed point based on the roughness, and perform plane point cloud growth based on the seed point to obtain a feature patch;

[0111] A reference patch determining unit is configured to randomly select a feature patch of an intermediate channel from the obtained feature patches as a reference patch.

[0112] Further, the angle error calculation module 30 comprises:

[0113] An error calculation unit is configured to calculate a channel angle error of each channel by taking the sum of squares of distances from feature points on a feature patch to reference points on a reference patch as a constraint condition;

[0114] A superposition unit is configured to superimpose the channel angle error of each channel with a theoretical error respectively, and determine a superposition result as an angle error calibration result of the corresponding channel.

[0115] Further, the error calculation unit comprises:

[0116] A feature coordinate calculation unit is configured to calculate feature coordinates of feature points on a feature patch based on original coordinates of laser foot points in a laser sensor original coordinate system, a rotation matrix, and coordinates of feature points in a scene coordinate system;

[0117] A first distance equation determining unit is configured to determine a first distance equation of feature points to a reference plane based on the feature coordinates and the normal vector of the feature points;

[0118] A second distance equation determining unit is configured to linearize the first distance equation using a Taylor expansion to obtain a second distance equation;

[0119] An error equation determining unit is configured to obtain an unknown coefficient matrix based on the second distance equation, and determine an error equation based on the unknown coefficient matrix and a distance of laser foot points to the reference plane;

[0120] An angle error determining unit is configured to determine an angle error of the corresponding channel based on the error equation and a least square method principle.

[0121] Further, the file obtaining module further comprises:

[0122] An obtaining unit is configured to control a UAV to fly along a preset flight path on a pre-arranged site, and obtain the point cloud data file, wherein the UAV is mounted with a multi-channel laser sensor to be calibrated and an inertial navigation device.

[0123] Further, the angle error calculation module further comprises:

[0124] A calibration unit is configured to calibrate each channel based on the angle error calibration result, thereby completing the calibration of the inter-channel angle error of the multi-channel laser sensor.

[0125] Referring to Figure 10 The present application also provides a device, the internal structure of which can be as shown in Figure 10 The device comprises a processor, a memory, a network interface and a database connected through a system bus. The processor is configured to provide computing and control capabilities. The memory of the device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium. The database of the device is configured to be used for the target recognition program of the device. The network interface of the device is configured to be used for communication with an external terminal through a network connection. The input device of the device is configured to receive signals input by an external device. The computer program is configured to be executed by the processor to implement a multi-channel laser sensor inter-channel angle error flight calibration method as described in the above embodiments.

[0126] Those skilled in the art can understand Figure 10 that the structure shown in the above

[0127] In addition, the present application also provides a computer readable storage medium comprising a multi-channel laser sensor inter-channel angle error flight calibration program, wherein the multi-channel laser sensor inter-channel angle error flight calibration program is configured to be executed by a processor to implement the steps of the multi-channel laser sensor inter-channel angle error flight calibration method as described in the above embodiments. It can be understood that the computer readable storage medium in the present embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0128] To sum up, the multi-channel laser sensor inter-channel angle error flight calibration method, device, equipment and computer readable storage medium provided in the embodiments of the present application calibrate and split the obtained point cloud data file to obtain a channel point cloud data file; determine a feature face sheet and a reference face sheet based on the channel point cloud data in the channel point cloud data file, wherein the reference face sheet is a feature face sheet corresponding to the middle channel; calculate the channel angle error of each channel according to the distance from the feature points on the feature face sheet to the reference points on the reference face sheet, and determine the angle error calibration result based on the channel angle error and the theoretical error. In this way, the angle error of each channel of the multi-channel laser sensor is calibrated, and the precision of the multi-channel laser sensor is improved.

[0129] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium provided by the present application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM) and the like.

[0130] It should be noted that in this document, the terms "comprising", "including", or any other variant thereof are intended to cover non-exclusive inclusions, so that processes, devices, articles or methods including a series of elements not only include those elements, but also include other elements not explicitly listed, or include elements inherent to such processes, devices, articles or methods. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0131] The above merely provides the preferred embodiments of the present application, and is not intended to limit the patent scope of the present application. Any equivalent structure or equivalent flowchart transformation based on the content of the present application specification and drawings, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method of flight calibration of inter-channel angle errors for a multi-channel laser sensor, characterized in that, The method comprises the following steps: Calibration and splitting of the obtained point cloud data file to obtain a channel point cloud data file; Determination of a feature patch and a reference patch based on channel point cloud data in the channel point cloud data file, wherein the reference patch is a feature patch corresponding to a middle channel; Calculation of a channel angle error of each channel according to a distance from a feature point on the feature patch to a reference point on the reference patch, and determination of an angle error calibration result based on the channel angle error and a theoretical error; The calibration and splitting of the obtained point cloud data file to obtain a channel point cloud data file comprises: Adjustment and calibration of the obtained original point cloud data to obtain an overall installation error of an IMU and a laser sensor body; Calibration of the original point cloud data based on the overall installation error to obtain calibrated point cloud data; Splitting of the calibrated point cloud data according to flight lines to obtain flight line point cloud data files, the number of which is consistent with the number of flight lines; Splitting of any one flight line point cloud data file according to scanning lines to obtain channel point cloud data files, the number of which is consistent with the number of channels; The calculation of a channel angle error of each channel according to a distance from a feature point on the feature patch to a reference point on the reference patch comprises: Taking the sum of squares of distances from feature points on the feature patch to reference points on the reference patch as a constraint condition, the channel angle error of each channel is calculated; The channel angle error of each channel is superimposed with a theoretical error respectively, and the superimposed result is determined as an angle error calibration result of the corresponding channel.

2. The method of claim 1, wherein The determination of a feature patch and a reference patch based on channel point cloud data in the channel point cloud data file comprises: Blocking of the channel point cloud data to obtain channel point cloud data blocks; Calculation of a normal vector and a roughness of each feature point in the channel point cloud data block; Determination of a seed point based on the roughness, and plane point cloud growth based on the seed point to obtain a feature patch; An arbitrary selection of a feature patch of a middle channel from the obtained feature patches as a reference patch.

3. The method of claim 1, wherein, The calculation of a channel angle error of each channel according to a distance from a feature point on the feature patch to a reference point on the reference patch comprises: Calculation of feature coordinates of a feature point on the feature patch based on original coordinates of a laser foot point in a laser sensor original coordinate system, a rotation matrix, and coordinates of the feature point in a scene coordinate system; Determination of a first distance equation of the feature point to the reference plane based on the feature coordinates and the normal vector of the feature point; Linearization of the first distance equation using a Taylor expansion to obtain a second distance equation; Obtaining of a coefficient matrix of unknowns based on the second distance equation, and determination of an error equation based on the coefficient matrix of unknowns and a distance of the laser foot point to the reference plane; Determination of an angle error of the corresponding channel based on the error equation and the least square method principle.

4. The method of claim 1, wherein Before the calibration and splitting of the obtained point cloud data file to obtain a channel point cloud data file, the method further comprises the following steps: The unmanned aerial vehicle flies along a preset flight path on a pre-arranged site to obtain the point cloud data file, wherein the unmanned aerial vehicle is mounted with a multi-channel laser sensor to be calibrated and an inertial navigation device.

5. The method of claim 1, wherein The method further comprises: based on the angle error calibration result, calibrating each channel to complete the calibration of the angle error between the channels of the multi-channel laser sensor.

6. A multi-channel laser sensor inter-channel angle error in-flight calibration device, characterized in that, The method comprises: a file obtaining module configured to calibrate and split the obtained point cloud data file to obtain a channel point cloud data file; a patch determining module configured to determine a feature patch and a reference patch based on the channel point cloud data in the channel point cloud data file, wherein the reference patch is a feature patch corresponding to an intermediate channel; an angle error calculating module configured to calculate the channel angle error of each channel according to the distance between the feature points on the feature patch and the reference points on the reference patch, and determine an angle error calibration result based on the channel angle error and a theoretical error; the file obtaining module is specifically configured to: adjust and calibrate the obtained original point cloud data to obtain the overall installation error of the IMU and the laser sensor body; based on the overall installation error, calibrate the original point cloud data to obtain calibrated point cloud data; split the calibrated point cloud data according to the flight path to obtain flight path point cloud data files, the number of the flight path point cloud data files being consistent with the number of the flight paths; split any one of the flight path point cloud data files according to the scanning lines to obtain channel point cloud data files, the number of the channel point cloud data files being consistent with the number of the channels; the angle error calculating module is specifically configured to: minimize the sum of the squares of the distances between the feature points on the feature patch and the reference points on the reference patch as a constraint condition to calculate the channel angle error of each channel; superimpose the channel angle error of each channel and the theoretical error respectively to determine the superimposed result as the angle error calibration result of the corresponding channel.

7. A multi-channel laser sensor inter-channel angle error flight calibration device, characterized by, The device comprises a memory, a processor, and a multi-channel laser sensor inter-channel angle error flight calibration program stored on the memory and executable on the processor, and the multi-channel laser sensor inter-channel angle error flight calibration program, when executed by the processor, implements the steps of the multi-channel laser sensor inter-channel angle error flight calibration method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a multi-channel laser sensor inter-channel angle error flight calibration program, and the multi-channel laser sensor inter-channel angle error flight calibration program, when executed by the processor, implements the steps of the multi-channel laser sensor inter-channel angle error flight calibration method according to any one of claims 1 to 5.

Images