Slope monitoring data calibration and three-dimensional point cloud matching method

Abstract

The invention discloses a slope monitoring data calibration and three-dimensional point cloud matching method, belongs to the field of monitoring data calibration, and aims to improve orientation errors caused by the fact that global satellite positioning and navigation system measurement endpoint errors and equipment length problems are not fully considered in existing slope radar measurement. The method solves the problem of position deviation of a monitoring result caused by the fact that the position of a deformation area is monitored, breaks through the limitation of insufficient precision when a traditional radar image is converted into geographic space coordinates, improves the accuracy and reliability of data, remarkably improves the position precision of slope radar monitoring data, and enables the position monitoring of the deformation area to be more accurate. The high-precision requirement of slope stability monitoring is met, reliable data support is provided for fine monitoring of slope stability, potential safety hazards can be found in time, and corresponding measures can be taken.

Description

Technical Field

[0001] This invention relates to the field of monitoring data calibration, and more specifically, to a method for calibrating slope monitoring data and matching it with a three-dimensional point cloud. Background Technology

[0002] Slope radar technology, as an application of ground-based synthetic aperture radar (GB-SAR) technology, effectively observes minute surface deformations using interferometric methods. In recent years, with the continuous improvement of open-pit mining development in China, the depth and scope of mining have expanded year by year, leading to increasingly prominent safety hazards such as slope collapses and landslides, which urgently require attention. Due to its small size, ease of deployment, and ability to perform long-term monitoring, slope radar equipment has been widely used in the field of slope stability monitoring.

[0003] Currently, slope radar typically determines the radar's orientation and center point coordinates by measuring the left and right endpoints of the radar, thereby converting the coordinates of the monitored deformation map or radar image into geospatial coordinates.

[0004] Existing research has not fully considered the errors inherent in the measurement endpoints of Global Navigation Satellite Systems (GNSS), and due to the relatively short length of radar equipment, the orientation calculated based on this endpoint will have a certain angular error. Because of this error, the positional error of radar images increases with distance when performing 3D matching or converting to geospatial coordinates, ultimately leading to a shift in the location of the deformation zone.

[0005] Therefore, this invention proposes a more accurate method for calculating radar orientation and achieves high-precision terrain matching between radar image correction and laser point cloud, thereby improving the positional accuracy of slope radar monitoring data and meeting the high-precision requirements for slope stability monitoring. Summary of the Invention

[0006] 1. Technical problems to be solved

[0007] In view of the problems existing in the prior art, the purpose of this invention is to provide a method for calibrating slope monitoring data and matching three-dimensional point clouds, which solves the problem of positional offset of monitoring results caused by radar equipment orientation measurement error, and realizes the matching of radar images and three-dimensional laser point clouds.

[0008] 2. Technical Solution

[0009] To solve the above problems, the present invention adopts the following technical solution.

[0010] A method for calibrating slope monitoring data and matching it with a 3D point cloud includes the following steps:

[0011] S1. Using the Global Positioning System (GPS), measure the radar measurement starting point L(X) at both ends of the slope radar equipment.L ,Y L H L The spatial coordinates of the radar measurement starting point R(X) R ,Y R H R The spatial coordinates of the slope radar equipment are obtained, and the center points O(X) of the two slope radar devices are obtained. O ,Y O H O The coordinates of the two slope radar devices, specifically the center point O, are calculated using the following formula:

[0012]

[0013] Then, the initial measurement orientation angle φ between the two slope radar devices is calculated using the following formula:

[0014]

[0015] Where X L Y L H L X R Y R H R These are the three-dimensional spatial coordinate parameters measured by the Global Positioning System (GPS).

[0016] S2. Install at least two corner reflectors on the measured slope surface. Use a global positioning and navigation system to measure the three-dimensional spatial coordinates of the corner reflectors as follows: {(X1,Y1,H1),…,(X... n ,Y n H n )}, where n is the number of corner reflectors, and n≥2, X1,Y1,H1,X n ,Y n H n These are the three-dimensional spatial coordinate parameters measured by the Global Positioning System (GPS).

[0017] S3. Turn on the radar equipment, mark the positions of all the corner reflectors in the radar image system, and obtain the set of two-dimensional coordinates of all the corner reflectors in the radar image system as: {(r1,θ1),…,(r n ,θ n )}, where n is the number of corner reflectors, r1, θ1, r n ,θ n These are the two-dimensional spatial coordinate parameters measured in the radar image system;

[0018] S4. Using a LiDAR laser point cloud system, obtain the set of three-dimensional coordinates of the laser point cloud at the slope surface as follows: Where m is the number of point clouds. These are the spatial coordinate parameters measured in the LiDAR laser point cloud system.

[0019] S5. Transform the three-dimensional coordinate set of the corner reflector described in step S2 and the three-dimensional coordinate set of the laser point cloud described in step S4 into the coordinate system of the radar image system with the center point O of the slope radar equipment as the origin. The transformation formulas are as follows:

[0020] x i =X n -X o

[0021] y i =Y n -Y o

[0022] h i =H n -H o

[0023]

[0024] The transformed three-dimensional coordinate set of the corner reflector is obtained as: {(x1,y1,h1),…,(x i ,y i ,h i The set of laser point cloud coordinates is as follows:

[0025] S6, Let point P(p) x ,p y Let O be a point on the unit circle centered at O ​​on the radar plane in the radar imaging system, and The direction is the actual orientation of the radar, and the angle θ of the corner reflector Q on the radar plane in the radar image system is obtained from the slope radar geometry. Q equal and The angle between the points is given, and the corner reflector Q is any one of n corner reflectors, and since the point P lies on the unit circle, it satisfies (p x ) 2 +(p y ) 2 =1, specifically we can get:

[0026]

[0027] Furthermore, this equation is satisfied for all n number of the aforementioned corner reflector devices;

[0028] Integrating the obtained system of equations, we get the error function E(p) x ,p yThe formula for ) is:

[0029]

[0030] The least squares method is used to iteratively optimize and minimize the error function, and a point on the unit circle is selected for the initial radar measurement in the direction φ. As the initial value for the iteration, specifically:

[0031]

[0032] Solving for the optimal parameter p x ,p y By minimizing the error function value, and combining the actual monitoring distance and the accuracy requirements of the monitoring position, it is determined whether the accuracy requirements are met after correction. If not, the number of corner reflectors needs to be increased and the correction process needs to be repeated.

[0033] S7. Calculate the slope distance between the randomly selected laser point cloud K coordinate and point O after transformation, and set it as... and The included angle is set as The set of slant ranges and included angles of the laser point cloud is obtained as follows: The specific calculation formula is as follows:

[0034]

[0035] S8. Let point j be a general point in the radar image. Calculate its weighted Euclidean distance to each point i in the laser point cloud. The specific calculation method is as follows:

[0036]

[0037] in θ j All measurements are in radians. For point j in each radar image, find the value that makes D... ij The smallest point cloud i, i.e.:

[0038]

[0039] Establish a matching index relationship set {(j,i)} between radar image point j and point cloud i. Based on the matching index relationship, assign the three-dimensional laser point cloud coordinates of the radar image point to the radar image point to complete the three-dimensional matching of radar image and point cloud.

[0040] Furthermore, in step S2, the placement of the corner reflectors satisfies the condition that the lines connecting any two corner reflectors and the radar center point O are not collinear, and that they are distributed at different heights on the slope surface.

[0041] Furthermore, the termination condition for the least squares iterative optimization in step S6 is: the difference between the error function values ​​of two adjacent iterations is less than 1 × 10⁻⁶. -6 Or the number of iterations reaches a preset threshold.

[0042] Furthermore, the weighting coefficient of the weighted Euclidean distance mentioned in step S8 is dynamically adjusted according to the slope. When the slope is greater than 30°, the angle difference weighting coefficient increases by 1.2-1.5 times.

[0043] 3. Beneficial effects

[0044] Compared with the prior art, the advantages of this invention are:

[0045] (1) This scheme improves the orientation error caused by insufficient consideration of the measurement endpoint error of the global satellite positioning and navigation system and the equipment length problem in the existing slope radar measurement, which leads to the position offset of the monitoring results. It breaks through the limitation of insufficient accuracy when converting traditional radar images to geospatial coordinates, improves the accuracy and reliability of the data, and significantly improves the position accuracy of slope radar monitoring data, making the position monitoring of deformation zone more accurate, meeting the high precision requirements of slope stability monitoring, providing reliable data support for the refined monitoring of slope stability, and helping to discover potential safety hazards in a timely manner and take corresponding measures.

[0046] (2) In this scheme, the coordinates of the two ends of the slope radar are accurately measured by the global satellite positioning and navigation system, and then the coordinates of the center point of the equipment and the initial orientation are determined to provide basic data for subsequent calibration. Furthermore, multiple corner reflectors are deployed on the slope to obtain their radar image coordinates and three-dimensional spatial coordinates, and the relevant coordinates are converted to the local coordinate system of the center point of the equipment for unified calculation and analysis, so as to accurately determine the equipment parameters.

[0047] (3) In this scheme, an error function is established based on the geometric relationship between the corner reflector and the radar center. The initial orientation measured by the global satellite positioning and navigation system is used as the initial value for iteration. The least squares method is used to iteratively solve the actual orientation, which effectively reduces the orientation angle error caused by the short length of the equipment. The correction accuracy is judged based on the minimized error function value and the actual monitoring requirements. If necessary, a corner reflector is added for recalibration to ensure the accuracy of the radar orientation and achieve the effect of optimizing the radar orientation calculation.

[0048] (4) In this scheme, the slant distance between the point cloud and the center of the equipment and the angle between the point cloud and the actual orientation are calculated. The weighted Euclidean distance between each radar image point and each point in the point cloud is calculated. The point cloud corresponding to the minimum distance is found, and a matching index relationship is established. At the same time, the point cloud coordinates are assigned to the radar image points according to the matching index, so as to achieve high-precision matching between the radar image and the three-dimensional laser point cloud. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the top-view geometry of the slope radar according to the present invention. Detailed Implementation

[0050] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0051] Example 1:

[0052] Please see Figure 1 A method for calibrating slope monitoring data and matching it with 3D point clouds is proposed, which achieves accurate calibration of slope monitoring data and high-precision matching between radar images and 3D point clouds. Its specific working principle is as follows:

[0053] I. Initial Radar Parameter Acquisition and Coordinate System Establishment

[0054] 1. Measurement of radar endpoint and center point coordinates

[0055] Using the Global Positioning System (GPS) to measure the three-dimensional spatial coordinates L(X) of the starting points L and R of the radar measurement at both ends of the slope. L ,Y L H L ) and R(X R ,Y R H R The coordinates of the radar center point O are calculated by averaging the coordinates of the two endpoints.

[0056] The central point O serves as the origin for subsequent coordinate transformations, providing a reference for a unified coordinate system.

[0057] 2. Initial orientation angle calculation: Based on the coordinate difference between L and R, the initial radar measurement orientation angle φ is calculated using the arctangent function, i.e. This angle reflects the initial measurement direction of the radar, providing iterative initial values ​​for subsequent calibration.

[0058] II. Establishment of Corner Reflector Calibration Reference

[0059] 1. Corner reflector placement and coordinate measurement

[0060] At least two corner reflectors (n≥2) are deployed on the slope surface, ensuring that the lines connecting any two corner reflectors to the radar center point O are not collinear and are distributed in different height areas to avoid geometric degradation and improve calibration accuracy. Then, the three-dimensional coordinates of each corner reflector are measured using the Global Positioning and Navigation System, forming a set {(X1,Y1,H1),…,(X... n ,Y n H n )}.

[0061] 2. Extraction of corner reflector coordinates from radar images

[0062] Turn on the radar equipment, mark the positions of each corner reflector in the radar image system, and obtain their two-dimensional polar coordinates, with distance r and angle θ, forming a set {(r1,θ1),...,(r...}}. n ,θ n This data forms a "ground truth value - radar observation" correspondence with the three-dimensional coordinates of the Global Positioning and Navigation System, providing constraints for calibration.

[0063] III. Coordinate System Transformation

[0064] The coordinates of the corner reflector and the laser point cloud are uniformly transformed to a local coordinate system with the radar center point O as the origin, eliminating absolute coordinate differences and facilitating geometric calculations. The transformation process is as follows:

[0065] 1. Corner reflector coordinate transformation:

[0066] For each corner reflector, calculate the coordinates relative to point O: x i =X n -X o y i =Y n -Y o h i =H n -H o This forms the transformed set {(x1,y1,h1),…,(x i ,y i ,h i )}.

[0067] 2. Laser point cloud coordinate transformation:

[0068] The set of three-dimensional coordinates of laser point clouds acquired by LiDAR Calculate the coordinates relative to point O: Form the transformed set

[0069] IV. Precise Radar Orientation Calibration

[0070] 1. Construction of the error function

[0071] Let the radar's true orientation be point P(p) on the unit circle. x ,p y ), and satisfy (p x ) 2 +(p y ) 2 =1, according to geometric relationships, the angle θ of the corner reflector Q in the radar image is... Q It should be equal to the angle between OQ and OP, from which the equation can be established:

[0072]

[0073] Based on the constraints of n corner reflectors, an error function (reflecting the deviation between observed and theoretical values) is constructed:

[0074]

[0075] 2. Least Squares Optimization

[0076] Point corresponding to the initial orientation φ and As the initial value for iteration, E(p) is minimized iteratively using the least squares method. x ,p y Solve for the optimal parameters (p) x ,p y The true orientation is given by the given direction, and the iteration terminates when the error difference between two consecutive iterations is 1 × 10⁻⁶. -6 The number of iterations may be preset to ensure convergence accuracy.

[0077] 3. Accuracy Verification and Adjustment

[0078] If minimizing the error meets the monitoring accuracy requirements, the calibration is complete; otherwise, increase the number of corner reflectors (enhance the constraints) and re-perform the calibration until the accuracy meets the requirements.

[0079] V. Matching Radar Imagery with Laser Point Clouds

[0080] 1. Polar coordinate transformation of laser point cloud

[0081] For the converted laser point cloud, calculate the slant distance of each point K relative to point O. and the angle between the actual orientation towards the OP The polar coordinate set that forms the point cloud This aligns it with the polar coordinate system of radar imagery.

[0082] 2. Weighted Euclidean Distance Matching

[0083] For a point j in a radar image, the coordinates (r) j ,θ j ), calculate its coordinates relative to the coordinates of each point i in the laser point cloud. Weighted Euclidean distance: The angle difference weight is dynamically adjusted according to the slope gradient. When the slope is greater than 30°, the angle difference weight increases by 1.2-1.5 times to compensate for the influence of angle error under steep terrain. This is achieved by finding the minimum D. ij For the corresponding point i, establish a matching index {(j,i)} between radar image point j and point cloud i.

[0084] 3. Assigning three-dimensional coordinates

[0085] Based on the matching index, the three-dimensional coordinates of the laser point cloud are assigned to the corresponding radar image points, thereby realizing the spatial association between the radar image and the three-dimensional point cloud and completing the three-dimensional matching.

[0086] This method eliminates radar orientation errors through corner reflector calibration, and combines coordinate transformation and weighted matching algorithms to achieve high-precision calibration of slope monitoring data and accurate fusion of radar images and three-dimensional point clouds, providing reliable spatial positioning support for slope stability monitoring.

[0087] Example 2:

[0088] Application in monitoring high and steep highway slopes

[0089] I. Scenario Background and Equipment Selection

[0090] This embodiment addresses a steep slope at kilometer marker K12+300 on a mountainous highway. The slope is approximately 40°, 80m high, and 150m long. Due to recent rainfall, the slope has experienced localized rockfalls. The method described in this invention is needed to calibrate radar monitoring data and match it with three-dimensional point clouds, providing precise spatial positioning support for slope stability assessment.

[0091] Equipment selection:

[0092] Global Positioning and Navigation Satellite System: The BeiDou-3 static positioning system is adopted, with one base station (stabilized rock mass 300m away from the slope) and the rover station using a Trimble R10 receiver. Positioning accuracy: horizontal ±(2mm+0.5ppm), elevation ±(5mm+0.5ppm), sampling frequency 1Hz.

[0093] Slope radar equipment: A certain type of ground-based synthetic aperture radar (GB-SAR) was selected, with a working frequency band of Ku band (17.2GHz), a ranging range of 50-500m, an angular resolution of 0.05°, a distance resolution of 0.3m, and a scanning cycle of 10 minutes / frame.

[0094] LiDAR laser point cloud system: Employs a terrestrial 3D laser scanner (FaroFocus S70), with a point cloud density ≥100 points / m². 2Distance measurement accuracy ±2mm@50m, scanning range 360° horizontally and 90° vertically, data format LAS1.2.

[0095] Corner reflectors: 15cm×15cm square metal corner reflectors (reflectivity ≥95%) are selected, and a total of 4 (n=4) are deployed. They are fixed to the slope rock mass by concrete pouring to ensure that the angle between the surface and the radar beam normal is ≤3°.

[0096] II. Specific Steps and Numerical Examples

[0097] 1. Radar initial parameter measurement (as described in step S1)

[0098] Measure the coordinates of the two ends of the radar:

[0099] Starting point L: (56789.123, 123456.789, 320.50) m (plane coordinates adopt Gauss-Kruger 3-degree zone projection, elevation is 1985 National Elevation Datum), starting point R: (56791.456, 123458.901, 320.62) m;

[0100] Calculate the coordinates of the center point O:

[0101] X O =(56789.123+56791.456) / 2=56790.2895m

[0102] Y O = (123456.789 + 123458.901) / 2 = 123457.845 m

[0103] H O = (320.50 + 320.62) / 2 = 320.56 m

[0104] Initial orientation angle φ:

[0105] φ=arctan[(56791.456-56789.123) / (123458.901-123456.789)]=arctan(2.333 / 2.112)≈48.2°

[0106] 2. Corner reflector placement and coordinate measurement (as described in steps S2-S3)

[0107] Deployment location:

[0108] Four corner reflectors are distributed at 20m intervals along the slope height, and the angle between any two points and point O is ≥35° (to avoid collinearity):

[0109] Q1 (bottom of slope): (56800.345, 123465.678, 310.20)m

[0110] Q2 (lower part of the slope): (56805.678, 123460.123, 330.50)m

[0111] Q3 (middle and upper part of the slope): (56810.901,123455.456,350.80)m

[0112] Q4 (Slope Top): (56815.234, 123450.789, 370.10)m

[0113] Radar image coordinates:

[0114] The two-dimensional polar coordinates of the corner reflector were obtained through radar image calibration:

[0115] Q1: (r = 75.3m, θ = 49.5°)

[0116] Q2: (r = 82.6m, θ = 46.2°)

[0117] Q3: (r = 89.8m, θ = 43.1°)

[0118] Q4: (r = 96.5m, θ = 40.3°)

[0119] 3. Coordinate transformation (as described in step S5)

[0120] Transform the corner reflector coordinates to a local coordinate system with O as the origin (x = X - X_O, y = Y - Y_O, h = H - H_O):

[0121] Q1: x=56800.345-56790.2895=10.0555m; y=123465.678-123457.845=7.833m;

[0122] h = 310.20 - 320.56 = -10.36m

[0123] Q2: x=56805.678-56790.2895=15.3885m; y=123460.123-123457.845=2.278m;

[0124] h = 330.50 - 320.56 = 9.94m.

[0125] 4. Radar orientation calibration (as described in step S6)

[0126] Error function construction: Based on the constraints of four corner reflectors, the error function E(p) is constructed. x ,py ) represents the square mean of the deviations of the four corner reflectors.

[0127] Iterative optimization:

[0128] Initial iteration point P (0) = (cos48.2°, sin48.2°)≈(0.667, 0.745), initial error E≈0.031;

[0129] After 3 iterations: P (3) = (0.672, 0.741), E≈0.0008;

[0130] After 5 iterations: P (5) = (0.673, 0.740), E≈0.00009 (satisfies the termination condition: the difference in error between adjacent iterations < 1 × 10⁻⁶). -6 The final true orientation angle is approximately 47.9°.

[0131] 5. Laser point cloud processing and matching (as described in steps S7-S8)

[0132] Point cloud polar coordinate transformation: Calculate the slope distance and angle of a point K in a point cloud (1.2 million points) acquired by LiDAR.

[0133] Transformed coordinates: x L =8.52m,y L =6.31m,h L = -8.20m

[0134] Slope distance

[0135] included angle

[0136] Weighted matching: For radar image point j (r = 70m, θ = 47.5°), calculate the weighted distance to the point cloud:

[0137] Point cloud i:R L =69.8m,θ L =47.3°

[0138] Angular difference (radians): 47.3° - 47.5° = -0.2° × π / 180 ≈ -0.0035 rad

[0139] Distance difference: 69.8 - 70 = -0.2 m

[0140] Weighted distance (Since the slope of 40° is greater than 30°, the angle difference weight is taken as 1.3, and the final D is calculated.) ij ≈0.28 m)

[0141] Matching result: The radar point is matched to point cloud i and assigned three-dimensional coordinates (56798.81, 123464.16, 312.36)m.

[0142] III. Accuracy Verification and Advantage Analysis

[0143] 1. Calibration accuracy:

[0144] Traditional method (orientation calculated only at endpoints): horizontal position error at a distance of 200m ≈ 0.8m;

[0145] The method of this invention (n=4 corner reflectors) reduces the distance error to 0.15m, meeting the accuracy requirements for highway slope monitoring (±0.3m).

[0146] 2. Matching results:

[0147] After weight adjustment, the point cloud matching accuracy in the steep slope area (40°) increased from 76% to 93%, which is significantly better than unweighted matching.

[0148] 3. Stability: During 10 consecutive days of monitoring, the coordinate deviation of the same monitoring point was ≤0.08m, which verified the long-term reliability of the method.

[0149] IV. Abnormal Handling Instructions

[0150] 1. Corner reflector obstruction: Q3 was obstructed due to rockfall on the slope. Its historical coordinates were confirmed by retrospective analysis of radar images. Combined with Q1, Q2, and Q4, the error was still controlled within 0.2m.

[0151] 2. LiDAR point cloud noise: For local noise points (deviation > 0.5m), radius filtering is used (points with less than 3 points within a radius of 0.3m are removed) to improve matching stability.

[0152] The specific data demonstration of the steep highway slope scenario verifies the applicability of the invention in complex terrain, supplements details such as steep slope weight adjustment and anomaly handling, and further supports the feasibility and rigor of the technical solution.

[0153] The above description is merely a preferred embodiment of the present invention; however, the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concepts, should be covered within the scope of protection of the present invention.

Claims

1. A method for calibrating slope monitoring data and matching it with a three-dimensional point cloud, characterized in that, Includes the following steps: S1. Using the Global Positioning System (GPS), measure the radar measurement starting point L(X) at both ends of the slope radar equipment. L ,Y L H L The spatial coordinates of the radar measurement starting point R(X) R ,Y R H R The spatial coordinates of the slope radar equipment are obtained, and the center points O(X) of the two slope radar devices are obtained. O ,Y O H O The coordinates of the two slope radar devices, specifically the center point O, are calculated using the following formula: Then, the initial measurement orientation angle φ between the two slope radar devices is calculated using the following formula: Where X L Y L H L X R Y R H R These are the three-dimensional spatial coordinate parameters measured by the Global Positioning System (GPS). S2. Install at least two corner reflectors on the measured slope surface. Use a global positioning and navigation system to measure the three-dimensional spatial coordinates of the corner reflectors as follows: {(X1,Y1,H1),…,(X... n ,Y n H n )}, where n is the number of corner reflectors, and n≥2, X1,Y1,H1,X n ,Y n H n These are the three-dimensional spatial coordinate parameters measured by the Global Positioning System (GPS). S3. Turn on the radar equipment, mark the positions of all the corner reflectors in the radar image system, and obtain the set of two-dimensional coordinates of all the corner reflectors in the radar image system as: {(r1,θ1),…,(r n ,θ n )}, where n is the number of corner reflectors, r1, θ1, r n ,θ n These are the two-dimensional spatial coordinate parameters measured in the radar image system; S4. Using a LiDAR laser point cloud system, obtain the set of three-dimensional coordinates of the laser point cloud at the slope surface as follows: Where m is the number of point clouds. These are the spatial coordinate parameters measured in the LiDAR laser point cloud system. S5. Transform the three-dimensional coordinate set of the corner reflector described in step S2 and the three-dimensional coordinate set of the laser point cloud described in step S4 into the coordinate system of the radar image system with the center point O of the slope radar equipment as the origin. The transformation formulas are as follows: x i =X n -X o and i =Y n -AND o h i =H n -H o The transformed three-dimensional coordinate set of the corner reflector is obtained as: {(x1,y1,h1),…,(x i ,y i ,h i The set of laser point cloud coordinates is as follows: S6, Let point P(p) x ,p y Let O be a point on the unit circle centered at O ​​on the radar plane in the radar imaging system, and The direction is the actual orientation of the radar, and the angle θ of the corner reflector Q on the radar plane in the radar image system is obtained from the slope radar geometry. Q equal and The angle between the points is given, and the corner reflector Q is any one of n corner reflectors, and since the point P lies on the unit circle, it satisfies (p x ) 2 +(p y ) 2 =1, specifically we can get: Furthermore, this equation is satisfied for all n number of the aforementioned corner reflector devices; Integrating the obtained system of equations, we get the error function E(p) x ,p y The formula for ) is: The least squares method is used to iteratively optimize and minimize the error function, and a point on the unit circle is selected for the initial radar measurement in the direction φ. As the initial value for the iteration, specifically: Solving for the optimal parameter p x ,p y By minimizing the error function value, and combining the actual monitoring distance and the accuracy requirements of the monitoring position, it is determined whether the accuracy requirements are met after correction. If not, the number of corner reflectors needs to be increased and the correction process needs to be repeated. S7. Calculate the slope distance between the randomly selected laser point cloud K coordinate and point O after transformation, and set it as... and The included angle is set as The set of slant ranges and included angles of the laser point cloud is obtained as follows: The specific calculation formula is as follows: S8. Let point j be a general point in the radar image. Calculate its weighted Euclidean distance to each point i in the laser point cloud. The specific calculation method is as follows: in θ j All measurements are in radians. For point j in each radar image, find the value that makes D... ij The smallest point cloud i, i.e.: Establish a matching index relationship set {(j,i)} between radar image point j and point cloud i. Based on the matching index relationship, assign the three-dimensional laser point cloud coordinates of the radar image point to the radar image point to complete the three-dimensional matching of radar image and point cloud.

2. The method for calibrating slope monitoring data and matching three-dimensional point clouds according to claim 1, characterized in that: The placement of the corner reflectors in step S2 satisfies the condition that the lines connecting any two corner reflectors and the radar center point O are not collinear, and that they are distributed at different heights on the slope surface.

3. The method for calibrating slope monitoring data and matching three-dimensional point clouds according to claim 1, characterized in that: The termination condition for the least squares iterative optimization in step S6 is: the difference between the error function values ​​of two adjacent iterations is less than 1 × 10⁻⁶. -6 Or the number of iterations reaches a preset threshold.

4. The method for calibrating slope monitoring data and matching three-dimensional point clouds according to claim 1, characterized in that: The weighting coefficient of the weighted Euclidean distance mentioned in step S8 is dynamically adjusted according to the slope. When the slope is greater than 30°, the angle difference weighting coefficient increases by 1.2-1.5 times.

Images