Unmanned aerial vehicle intelligent inspection method and system based on FAST site dangerous rock

Abstract

The invention discloses an unmanned aerial vehicle intelligent routing inspection method and system based on FAST site dangerous rock, and the method comprises the steps: S1, carrying out the routing inspection through an unmanned aerial vehicle, carrying out the point cloud down-sampling, obtaining the radar data of a rock wall near the FAST site, and obtaining the point cloud data of the dangerous rock; s2, determining a back wall area, a back wall inclination angle, a maximum height difference, spatial distribution, a dangerous rock volume and a fracture attenuation coefficient; s3, setting a value interval for the dangerous rock mountain target, judging whether the value interval meets conditions or not according to the back wall area characteristic, the back wall inclination angle characteristic and the maximum height difference characteristic, and if the conditions are met, determining that the dangerous rock mountain target is a dangerous rock body; and S4, calculating the stability coefficient R of the dangerous rock, and carrying out qualitative evaluation on the stable state of the dangerous rock. According to the method, the unmanned aerial vehicle carrying the laser radar is used for collecting the FAST station site dangerous rock data, the dangerous rock is detected by adopting a multi-source data fusion dangerous rock feature extraction method, and the method has important significance for stable and reliable operation and scientific achievement output of FAST in national large science engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of astronomical observation and equipment maintenance, and in particular to an intelligent inspection method and system of unmanned aerial vehicle (UAV) based on dangerous rocks at the FAST site. Background Art

[0002] The Five-hundred-meter Aperture Spherical radio Telescope (FAST), built on a natural karst depression in Guizhou Province, is the world's largest single-aperture radio telescope, enabling high-precision astronomical observations over a large sky area. It pioneered a new model for giant telescope construction, surpassing the 100-meter engineering limit of traditional telescopes. Its enormous scale, ultra-high precision requirements, and unique operating methods created unprecedented technical challenges.

[0003] The FAST site covers a large area and is located in the Dawodang depression in Jinke Village, Kedu Town, Pingtang County, Qiannan Prefecture, Guizhou Province. It is a karst landform with a rugged surface and numerous rocks. The surface often has stone buds, stone forests, peak forests, karst gullies, funnels, sinkholes, karst depressions and other depressions. There are 5 higher peaks around the FAST site. The highest peak is located on the southeast side of the depression, with a peak elevation of 1201.20m. The maximum terrain height difference is 360.30m.

[0004] The eastern steep slopes and cliffs of the FAST site are widely distributed with structures composed of multiple groups of rock structure surfaces. Under the influence of gravity, earthquakes, water bodies, weathering and other inducing factors, they change into unstable, understable or extreme equilibrium dangerous rock masses. They exist on high and steep slopes and cliffs and collapse due to instability and movement. The steep terrain is the geomorphological feature of the development of dangerous rocks. The main modes of their destruction are sliding dangerous rocks and toppling dangerous rocks.

[0005] The collapse and fall of unstable rock masses in karst terrain is characterized by suddenness, high velocity, and unpredictability. Under the weight of the rock mass, fractures in the main structural surface, fissure water pressure, or extreme weather conditions, the stability of individual rock masses can be compromised. Movement on the slope can also affect other locally unstable rock masses. These rock masses can suddenly accelerate down the steep slopes surrounding the site, collide with the main structure, and cause friction that damages the steel structure's coating, accelerating the corrosion of metal exposed to air. These collisions can also directly cause irreparable damage to components, such as deformation, dents, fractures, and the loss of reflective panels. These damages pose a serious threat to the safe and stable operation of FAST. Every year, rockfalls collide with telescope equipment and facilities, causing incalculable damage and impacting the telescope's operational safety, observation time, and even the output of research.

[0006] Dangerous rocks at the FAST site are mainly distributed in the steep rock areas on the southeast and northwest sides. The rock types are dolomitic limestone and argillaceous limestone. According to investigations, there are ten major areas prone to dangerous rock masses at the FAST site. Dangerous rock masses may become unstable at any time, which is a common geological hazard at the FAST site. At present, the inspection of dangerous rocks at the site still uses traditional manual climbing visual inspections. This has problems such as low efficiency, many safety hazards, difficult climbing on steep cliffs, and lack of data assets. It is urgent to explore a new inspection method that is safe, efficient, reliable and convenient. Therefore, the use of an efficient and intelligent FAST site dangerous rock identification device to detect the volume of dangerous rock and its morphological characteristics, size and collapse development direction is the technical key to solving the impact of dangerous rock masses on the FAST project. Summary of the Invention

[0007] To address the challenges of the existing technology, the present invention aims to provide a method for intelligent drone inspection of dangerous rocks at the FAST site. This dangerous rock identification device is fast, efficient, cost-effective, and can be controlled from the ground. Another object of the present invention is to provide a drone intelligent inspection system for dangerous rocks at the FAST site that implements the aforementioned method.

[0008] To achieve the above objectives, the present invention provides an intelligent inspection method for dangerous rocks at the FAST site using a drone, comprising:

[0009] S1. Use drone inspections to perform point cloud downsampling, obtain radar data of rock walls near the FAST site, and obtain point cloud data of dangerous rocks;

[0010] S2. Determine the back wall area, back wall inclination, maximum height difference, spatial distribution, dangerous rock volume, and fracture attenuation coefficient;

[0011] S3. Set a value range for the dangerous rock mass target and determine whether it meets the conditions based on the back wall area characteristics, back wall inclination characteristics, and maximum height difference characteristics. If the conditions are met, it is determined to be a dangerous rock mass;

[0012] S4. Set the dangerous rock stability coefficient R to perform a qualitative evaluation of the stability state of the dangerous rock. The calculation formula of the stability coefficient R is:

[0013]

[0014] Where α is the crack attenuation coefficient, ρ is the density, V is the volume of dangerous rock, A is the area of the back wall, and θ is the inclination angle of the back wall. The greater the mass per unit area and the larger the inclination angle, the smaller the stability coefficient R and the worse the target stability.

[0015] Furthermore, for candidate targets with cracks on the back-avoidance surface and the rock mass body, α < 1, their stability will be further reduced.

[0016] Furthermore, the reconstructed point cloud of each candidate dangerous rock mountain target collected by the lidar is an irregular three-dimensional surface, and the least squares method is used to calculate the fitting plane for each candidate dangerous rock mountain target.

[0017] Furthermore, the projection of the outer contour of the candidate dangerous rock mountain target on the fitting plane is used as the boundary of the fitting plane, and the fitting plane is defined as the back wall plane of the object, and the area of the fitting plane is defined as the back wall area of the candidate dangerous rock mountain target.

[0018] Furthermore, the projection area of the candidate dangerous rock massif on the fitting plane is divided into several grids of equal size. The area of each grid cell within the corresponding boundary is calculated and finally multiplied by the number of grids to obtain the back wall area of the candidate object. The calculation formula is:

[0019] A=nab

[0020] Where n, a, and b are the number of subdivided grids and the length and width of the corresponding single minimum grid, respectively.

[0021] Furthermore, the angle between the back wall plane and the horizontal plane is the back wall inclination angle, denoted as θ, the horizontal plane normal vector is denoted as v, and the back wall plane normal vector is denoted as n. The back wall inclination angle of the candidate object is calculated as follows:

[0022]

[0023] Furthermore, the maximum vertical distance between the back wall plane and the corresponding grid surface is defined as the maximum height difference. The normal vector of the back wall plane is calculated, and then the normal vector is slid one by one on the back wall plane subdivision unit cells. The intersection of the normal vector and the midpoint of the i-th grid is recorded as A i The intersection point with the irregular surface of the dangerous rock mass is marked as B i , the maximum height difference calculation formula of the candidate object is:

[0024] ΔH=max{|A1B1|,|A2B2|···|A i B i |···}.

[0025] Furthermore, the take-off position of the UAV is taken as the coordinate origin, and the geometric center point P of the dangerous rock body annotation box is taken. The position coordinates of point P in the world coordinate system are the coordinates of the dangerous rock body object in space distribution, which can be expressed as:

[0026] P=(x0,y0,z0).

[0027] Furthermore, using an approximate calculation method, three points are taken from the point cloud 3D surface to generate several plane triangle meshes, which are connected to the center of the body to form a cone. The volume of each cone is V i, calculate the volume of all cones and add them up to define the volume of dangerous rock mass:

[0028] V=∑ i V i .

[0029] Furthermore, the cracks inside the rock are simplified, and the distance between the two extreme points where the normal vector has a non-uniform gradient change is defined as the crack opening d, and the angle between the normal vectors of the two points inside the crack is defined as β. Then the attenuation coefficient α is expressed as:

[0030] α=sigmoid[(β-π)lgd].

[0031] A drone intelligent inspection system based on dangerous rocks at the FAST site is provided, which is used to implement the above-mentioned drone intelligent inspection method based on dangerous rocks at the FAST site.

[0032] The technical effects of the present invention are:

[0033] (1) Using a drone equipped with a lidar to collect dangerous rock data at the FAST site, and using a multi-source data fusion dangerous rock feature extraction method to detect dangerous rocks, which is of great significance to the stable and reliable operation of the national large-scale scientific project FAST;

[0034] (2) Technologies such as the feature extraction algorithm based on true color 3D point cloud and the calculation method of dangerous rock feature parameters based on geometric features have certain innovative value;

[0035] (3) The technical achievements of the present invention can not only serve large scientific projects and improve the application capabilities of the model in different scenarios, but can also be applied to other similar projects. It is of great significance to improve the efficiency of dangerous rock inspection, reduce the risk of manual inspection and form data assets. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic diagram of the dangerous rock BBox;

[0037] Figure 2 Rendering for the Dangerous Rock target;

[0038] Figure 3 Schematic diagram of the fitted plane and horizontal plane;

[0039] Figure 4 Schematic diagram of the calculation model for the back wall area and back wall inclination;

[0040] Figure 5 This is a schematic diagram of the maximum height difference calculation model;

[0041] Figure 6 This is a schematic diagram of the spatial distribution of dangerous rocks;

[0042] Figure 7This is a schematic diagram of the dangerous rock mass volume calculation model;

[0043] Figure 8 Schematic diagram of the triangulation model;

[0044] Figure 9 Schematic diagram of the new triangle mesh model;

[0045] Figure 10 It is a schematic diagram of the Delaunay triangulation model;

[0046] Figure 11 This is a simplified model diagram of the crack. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solution of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0049] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0050] The following combination Figures 1-11 The specific embodiments of the present invention are described in detail. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0051] Dangerous rock masses refer to unstable rock masses and their combinations, located on steep cliffs or slopes and cut by multiple rock mass structural planes. Isolated dangerous rock masses on steep slopes appear as highly protruding blocks with distinct external edges or partial overhangs. Data processing and annotation are subject to subjective factors. The human eye cannot quantitatively judge the stability of rock masses, and some stable rock targets may be classified as candidate dangerous rock masses. Furthermore, 3D point cloud detection algorithms identify possible candidate dangerous rock masses by learning their geometric features, but they also detect some non-dangerous rock mass protrusions and objects in depressions (such as dead trees, vegetation, and potholes). The number of candidate dangerous rock masses exceeds the actual number of unstable dangerous rock masses, resulting in a certain degree of false detection (negative samples predicted as positive) while maintaining a high recall rate. Therefore, it is necessary to design a feature filter based on backwall area, backwall inclination, maximum height difference, dangerous rock volume, and fracture attenuation coefficient to select the final dangerous rock mass targets. The present invention utilizes drone inspections to perform point cloud downsampling, obtains rock wall radar data near the FAST site, and obtains dangerous rock point cloud data.

[0052] 1.1 Back wall area:

[0053] Since the laser radar collects the reflection intensity of the surface of the illuminated object, the point cloud reconstructed by each candidate dangerous rock body is an irregular three-dimensional surface. Therefore, the least squares method is used to calculate a fitting plane for each candidate dangerous rock body. The projection of the outer contour of the candidate dangerous rock body on the plane is used as the boundary of the plane, and the plane is defined as the back wall plane of the object. The area of the plane is defined as the back wall area of the candidate dangerous rock body, such as Figure 1 、 Figure 2 、 Figure 3 shown.

[0054] The least squares method (also known as the least squares method) is a mathematical optimization technique. It finds the best function matching data by minimizing the sum of squared errors. Least squares can be used to easily obtain unknown data while minimizing the sum of squared errors between the obtained data and the actual data. In the field of imaging, the least squares method is often used for line fitting, curve fitting, and plane fitting.

[0055] For a plane, its equation can be expressed as z=ax+by+c. For a series of point cloud data {(x,y,z)|(x,y,z)∈(x i ,y i ,z i ), i=0,1,2,...,n-1}, by fitting the plane with discrete points, we need to find a plane (z=ax+by+c) such that the sum of the distances from the plane to each point cloud is S=∑(ax i +by i +cz i )2 Minimum, that is to say, a set of a, b, c is required so that for the existing discrete points, the value of S is minimized. Then formula (1) must be satisfied, and the partial derivative of the S value with respect to a, b, and c is zero. After expansion, formula (2) is obtained. Formula (2) is written in matrix form, such as formula (3). The matrix form is conducive to improving the calculation speed of the algorithm.

[0056]

[0057] The calculation formulas for constants a, b, and c are derived from the above calculation formulas. Using Cramer's rule, the expressions of a, b, and c are obtained as formulas (4), (5), and (6):

[0058]

[0059]

[0060] The least squares plane can be obtained: z=ax+by+c.

[0061] Then, the projection area of the candidate dangerous rock mass on the plane is divided into several grids of equal size, and the area of each divided cell within the corresponding boundary is calculated; finally, the area of the back wall of the candidate object is obtained by multiplying it by the number of grids. The calculation formula is:

[0062] A=nab (7)

[0063] Where n, a, and b are the number of subdivided grids and the length and width of the corresponding single minimum grid, respectively.

[0064] 1.2 Back wall inclination angle:

[0065] The angle between the back wall plane and the horizontal plane is defined as the back wall inclination angle, denoted as θ, such as Figure 4 As shown. Let the horizontal plane normal vector be v, the back wall plane normal vector be n, and the back wall inclination angle of the candidate object be calculated as:

[0066]

[0067] 1.3 Maximum height difference:

[0068] The maximum vertical distance between the back wall plane and the corresponding grid surface is defined as the maximum height difference, e.g. Figure 5 As shown. Calculate the normal vector of the back wall plane, then slide the normal vector one by one on the back wall plane subdivision unit cell, and record the intersection of the normal vector and the midpoint of the i-th grid as A i The intersection point with the irregular surface of the dangerous rock mass is marked as B i , the maximum height difference calculation formula of the candidate object is:

[0069] ΔH=max{|A1B1|,|A2B2|···|A i B i |···} (9)

[0070] After obtaining the maximum height difference from the above formula, find the vector A corresponding to the maximum height difference i B i The angle α between the ΔH and the horizontal plane normal vector is calculated as shown in Formula 6-24. If 0°<α<90°, the sign of ΔH is positive, otherwise it is negative.

[0071]

[0072] 1.4 Spatial distribution:

[0073] The take-off position of the UAV is the coordinate origin, and the geometric center point P of the dangerous rock mass mark box is taken as follows: Figure 6 As shown in the figure, the position coordinates of point P in the world coordinate system are the coordinates of the dangerous rock mass object in space distribution, which can be expressed as:

[0074] P=(x0,y0,z0) (11)

[0075] 1.5 Volume of dangerous rock:

[0076] Using the approximate calculation method, three points are taken on the point cloud 3D surface to generate several plane triangle meshes, which are connected to the center of the body to form a cone. The calculation model is as follows: Figure 7 As shown, the volume of each cone is Vi. The volumes of all cones are calculated and accumulated to define the formula for calculating the volume of dangerous rock mass:

[0077] V=∑ i V i (12)

[0078] In order to calculate the volume Vi of a single cone, an irregular triangulated network (TIN) needs to be constructed to calculate the volume of dangerous rocks. For point cloud data, due to its huge amount of data, its ordinary network construction method such as Delaunay triangulation will take a lot of time, so a calculation method based on Bowyer-Watson is adopted. The Bowyer-WYatson algorithm is specifically as follows: first, according to the point cloud coordinate distribution, a super triangle containing the entire calculation area is defined, and a triangle ABC is taken from the triangle set, and its vertices are set to A(x1, y1), B(x2, y2), C(x3, y3), and the circumscribed circle of the triangle is calculated. Then, a point P1 is inserted, and all triangles whose circumscribed circles contain this point are found. The cavities formed by these triangles are deleted. The triangulation model is as follows: Figure 8 As shown, the formula for calculating the center and radius of the circumscribed circle is:

[0079]

[0080]

[0081] Connect the new points to the cavity nodes to form a new triangular mesh; update the data structure and fill the deleted triangle data with the newly generated triangle data; return to the step of calculating the circumscribed circle until all points are interpolated. Figure 9 Figure 2 shows a schematic diagram of a new triangular mesh model, where Figure b shows a schematic diagram of inserting the second point.

[0082] Take three points as an example to construct the point cloud data network, sort the points in ascending order along the X-axis, generate super triangles, insert points and update the triangulation network. When the interpolation starts, first use the number of coordinate points to control the number of times of the outermost loop, and loop all triangles internally. Before the inner loop, instantiate an edge class array to store the edges of all decomposed triangles. The operation inside the loop first determines whether the triangle is a decomposed triangle. If it is a decomposed triangle, proceed to the next loop. If it is not a decomposed triangle, then determine whether it is a Delaunay triangle. If it is a Delaunay triangle, jump out of the loop. If it is not a Delaunay triangle, then determine the positional relationship between the interpolation point and the triangle to determine whether to decompose the triangle. If decomposed, store the edges. After the inner loop is completed, delete the duplicate edges to form a cavity polygon and then generate a new triangle; insert the second point to determine whether the point is on the right side of the circumscribed circle (to determine whether the triangle is a Delaunay triangle); insert the second point to generate a new triangle; repeat the above steps to finally generate a Delaunay triangulation network, as shown in the schematic diagram. Figure 10 shown.

[0083] When calculating the volume of an irregular triangulated network, set the reference elevation to h0, take a triangle from the irregular triangle set, set its vertices to A(x1,y1), B(x2,y2), and C(x3,y3), and the volume can be calculated from the vertex coordinates and elevation data.

[0084] 1.6 Crack attenuation coefficient:

[0085] Simplify the cracks inside the rock into Figure 11 In the model shown, the distance between the two extreme points where the normal vector has a non-uniform gradient change is defined as the crack opening d, and the angle between the normal vectors of two points inside the crack is defined as β. The attenuation coefficient α is expressed as:

[0086] α=sigmoid[(β-π)lgd] (16)

[0087] Criteria for evaluating the stability of dangerous rock masses: Under the influence of rainfall, earthquakes, and weathering, cracks in dangerous rock masses and differential weathering can further intensify, creating the potential for slippage or collapse along fractured surfaces. Furthermore, under the combined influence of geological structures, joints and fissures, and surface water, weathered rock masses can partially penetrate concave cavities. Tensile cracks may develop at the top of these cavities, making them susceptible to breakage, slippage, or collapse along weaker surfaces (lower strength surfaces) under the effects of their own weight.

[0088] The present invention uses drones and laser radar to scan steep rock walls that are difficult to reach manually, and obtains the geometric characteristics and spatial distribution characteristics of dangerous rock bodies. From the perspective of the geometric characteristics of dangerous rock bodies, stability indicators and evaluation methods can be established. 2 A simple value range is set for dangerous rock mass targets of the order of area. The conditions are judged based on the area characteristics, inclination characteristics and maximum height difference characteristics. If the conditions are met, it can be identified as a dangerous rock mass, and a dangerous rock mass evaluation method is established. When collecting data this time, the sampling distance of the lidar was about 10m, and the dangerous rock mass targets on the FAST site rock wall photographed were smaller and more detailed. Therefore, considering the different contributions of the back wall area, volume and back wall inclination, the dangerous rock stability coefficient R was set to qualitatively evaluate the stability of the dangerous rock. The calculation formula of the stability coefficient R is:

[0089]

[0090] Where α is the crack attenuation coefficient, ρ is the density (assuming the same material, the density is the same), V is the volume of the dangerous rock, A is the area of the back wall, and θ is the back wall inclination angle. The greater the mass per unit area (the farther the center of mass is from the back wall plane) and the larger the inclination angle, the smaller the stability coefficient R and the worse the target stability. Furthermore, for candidate targets with cracks between the back surface and the rock mass, if α < 1, their stability will be further reduced.

[0091] The instability mode of dangerous rocks divides dangerous rocks into three categories: sliding dangerous rocks, tipping dangerous rocks and falling dangerous rocks. Sliding dangerous rocks: The dangerous rock body is exposed to the air on three sides (triangular prism, square prism, irregular body), closely attached to the parent rock, and unloading cracks are developed at the rear and bottom of the block. Under the influence of factors such as rainfall, the unloading cracks are connected and sliding failure is likely to occur. The sliding mechanism is that the anti-sliding force of the base of the dangerous rock is reduced under the action of external forces, and the ability to bear the gravity of the upper block is weakened, resulting in sliding failure of the dangerous rock block along the shear surface. Tipping dangerous rocks: Steeply inclined broken blocks and blocky rock masses bend under the potential energy of gravity, and bend and crack towards the air like a cantilever beam, and develop to the point where the root rock mass is fractured or even crushed, and then rotation and tipping failure occur.

[0092] The dangerous rock mass is primarily controlled by layers, fault planes, and structural planes, with well-developed unloading cracks at the trailing edge. The combination of these structural planes is extremely detrimental to the stability of the dangerous rock mass. Qualitative analysis indicates that the overall dangerous rock mass slope is generally stable, while some parts of the dangerous rock mass are unstable. A comprehensive analysis of the distribution, spatial geometry, structural characteristics, deformation characteristics, weak base, and influencing factors of the dangerous rock mass revealed that the instability mode of the dangerous rock mass at the FAST site is primarily toppling, with no large-scale sliding dangerous rock mass forming. A small amount of falling dangerous rock mass exists locally.

[0093] Detection and evaluation results: According to the stability coefficient calculation formula, the stability coefficients of dozens of candidate dangerous rock bodies in the dangerous rock area of the FAST site were calculated. The 1H Xiaowodang area was selected as the detection location to test the recognition rate of dangerous rock bodies during the intelligent inspection of the UAV. The detailed calculation results are shown in Table 1, and the candidate dangerous rock bodies are arranged in order of stability from small to large.

[0094] Table 1: Calculation table of stability coefficients of 9 representative candidate dangerous rock masses in the 1H Xiaowodang-1 area

[0095] Dangerous rock mass number Back wall area Back wall inclination Maximum height difference volume Stability coefficient R Steady state evaluation grade WY2 5.141 72.582 0.884 2.414 1.367 Instability Level 2 WY13 3.49 12.675 0.706 5.676 1.403 Instability Level 2 WY24 6.206 45.833 0.941 4.077 1.948 Instability Level 2 WY14 4.598 16.229 0.439 4.284 2.303 basically stable Level 3 WY26 0.379 52.153 0.411 0.159 2.733 basically stable Level 3 WY23 5.3 23.54 0.598 3.518 2.845 basically stable Level 3 WY10 4.438 39.445 0.506 1.937 3.279 basically stable Level 3 WY18 2.226 26.336 0.447 0.893 4.484 Stablize Level 4 WY11 4.021 24.307 0.281 1.104 6.688 Stablize Level 4

[0096] From the analysis of visual effects and actual stability coefficients in Table 1, it can be seen that according to the dangerous rock stability evaluation standard, WY2 and WY13 are unstable dangerous rock bodies, while the others are stable or basically stable. The on-site confirmation by geotechnical engineers is consistent with the results of the intelligent identification technology of the present invention, indicating that the technical results of the present invention are reliable and effective. At the same time, this technical result was also used to evaluate the stability of 105 candidate dangerous rock bodies in the dangerous rock area of the FAST site. After statistical comparison, the accuracy rate of dangerous rock identification was calculated to be 83.81%.

[0097] The present invention uses a laser radar-equipped drone to collect dangerous rock data at the FAST site and uses a multi-source data fusion dangerous rock feature extraction method to detect dangerous rocks. This method is of great significance to the stable and reliable operation of the national large-scale scientific project FAST and the output of scientific results.

Claims

1. A UAV intelligent inspection method based on dangerous rocks at the FAST site, characterized by: The method includes: S1. Use drone inspections to perform point cloud downsampling, obtain radar data of rock walls near the FAST site, and obtain point cloud data of dangerous rocks; S2. Determine the back wall area, back wall inclination, maximum height difference, spatial distribution, dangerous rock volume, and fracture attenuation coefficient; S3. Set a value range for the dangerous rock mass target and determine whether it meets the conditions based on the back wall area characteristics, back wall inclination characteristics, and maximum height difference characteristics. If the conditions are met, it is determined to be a dangerous rock mass; S4. Set the dangerous rock stability coefficient R to perform a qualitative evaluation of the stability state of the dangerous rock. The calculation formula of the stability coefficient R is: Among them, α is the crack attenuation coefficient, ρ is the density, V is the volume of dangerous rock, A is the area of the back wall, and θ is the inclination angle of the back wall. The greater the mass per unit area and the larger the inclination angle, the smaller the stability coefficient R and the worse the target stability.

2. The UAV intelligent inspection method based on dangerous rocks at the FAST site according to claim 1 is characterized in that: For each candidate dangerous rock mountain target collected by LiDAR, the reconstructed point cloud is an irregular three-dimensional surface, and the least squares method is used to calculate the fitting plane for each candidate dangerous rock mountain target.

3. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 2 is characterized in that: The projection of the outer contour of the candidate dangerous rock mountain target on the fitting plane is used as the boundary of the fitting plane, and the fitting plane is defined as the back wall plane of the object, and the area of the fitting plane is defined as the back wall area of the candidate dangerous rock mountain target.

4. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: The projection area of the candidate dangerous rock massif on the fitting plane is divided into several grids of equal size. The area of each grid cell within the corresponding boundary is calculated and finally multiplied by the number of grids to obtain the back wall area of the candidate object. The calculation formula is: A=nab Where n, a, and b are the number of subdivided grids and the length and width of the corresponding single minimum grid, respectively.

5. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: The angle between the back wall plane and the horizontal plane is the back wall inclination angle, denoted as θ, the horizontal plane normal vector is denoted as v, and the back wall plane normal vector is denoted as n. The back wall inclination angle of the candidate object is calculated as follows:

6. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: The maximum vertical distance between the back wall plane and the corresponding grid surface is defined as the maximum height difference. The normal vector of the back wall plane is calculated, and then the normal vector is slid one by one on the back wall plane subdivision unit cell. The intersection of the normal vector and the midpoint of the i-th grid is recorded as A. i The intersection point with the irregular surface of the dangerous rock mass is marked as B i , the maximum height difference calculation formula of the candidate object is: ΔH=max{|A1B1|,|A2B2|…|A i B i |…}。 7. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: The take-off position of the UAV is the coordinate origin. The geometric center point P of the dangerous rock mass annotation box is taken. The position coordinates of point P in the world coordinate system are the coordinates of the dangerous rock mass object in space, which can be expressed as: P=(x0,y0,z0).

8. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: Using an approximate calculation method, three points are taken from the point cloud 3D surface to generate several plane triangle meshes. The plane triangle meshes are connected to the center of the body to form a cone. The volume of each cone is V i , calculate the volume of all cones and add them up to define the volume of dangerous rock mass: V=∑ i V i 。 9. The method for intelligent inspection of dangerous rocks at the FAST site by using unmanned aerial vehicles according to claim 1 is characterized in that: Simplifying the cracks inside the rock, defining the distance between the two extreme points where the normal vector has a non-uniform gradient change as the crack opening d, and defining the angle between the normal vectors of the two points inside the crack as β, the attenuation coefficient α is expressed as: α=sigmoid[(β-π)lgd].

10. An intelligent drone inspection system for dangerous rocks at the FAST site, characterized by: The system is used to implement the UAV intelligent inspection method based on dangerous rocks at the FAST site as described in any one of claims 1-9.

Images