An open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud

Abstract

The application introduces an open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud, comprising the following steps: step S1, obtaining open-pit mine point cloud data: obtaining open-pit mine image data based on unmanned aerial vehicle oblique photogrammetry, importing into Context Capture, and then generating point cloud through image preprocessing, aerial triangulation encryption processing and multi-view image dense matching processing; step S2, open-pit mine point cloud data preprocessing: constructing neighborhood relationship based on Kd-Tree, realizing point cloud simplification based on geometric sampling, and simplifying data while retaining features; step S3, open-pit mine step element extraction: extracting slope surface according to neighborhood characteristics, selecting plane to manually divide slope surface, obtaining a plurality of step slope surfaces through density clustering processing of slope surface, and extracting step line; step S4, measuring step structure: calculating step height and step slope angle according to the extracted step element. Through the above design, the application can quickly and accurately measure the step structure, effectively reducing the open-pit mine supervision cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mine mapping, and particularly relates to a method for measuring open-pit mine step structure based on unmanned aerial vehicle image matching point cloud. BACKGROUND

[0002] Mineral resources are an important material basis for social production development, and rational development and comprehensive utilization of mineral resources is an inevitable requirement for modernization construction. Open-pit mining, as a low-cost, high-yield and high-efficiency production method, has been widely used in domestic and foreign mine engineering. Open-pit mines usually adopt a top-down and step-layered mining method. The step is the smallest mining unit in the vertical direction and is the operation site for tasks such as perforation, blasting and shovel loading, which directly affects the production scale, economic benefits and slope stability of the mine. Open-pit mining will form many high and steep slopes, and frequent construction operations such as blasting and stripping will cause damage to the step structure surface and deformation of the rock mass, eventually causing slope instability, deformation and other geological accidents, which seriously threatens the safety of personnel and property in the mine area.

[0003] Safe mining has become an extremely important part of mining engineering research, and runs through the whole process of open-pit mining, which is the prerequisite for the orderly production of open-pit mines. The shaping of the step structure in open-pit mining affects the structural stability and safety of the open-pit mine. If not effectively measured and controlled, it is easy to cause structural imbalance, which eventually leads to a series of geological disasters such as collapse, landslide and mudflow.

[0004] The traditional method uses surveying instruments such as level, total station and GNSS receiver to measure the step structure, which has high cost, low efficiency and cannot guarantee the safety of surveying personnel. Therefore, it is necessary to extract and measure the step structure by high-tech surveying technology. Continuous monitoring of the step structure during the mining process can effectively guide the production operation of the mining area and avoid imbalance. For example, Chinese patent CN112634389A discloses an open-pit mine plan drawing method, device and server based on a UAV, which comprises: obtaining a scene photo of a target area; importing a current month orthographic image, radar point cloud and last month step line; displaying a radar point cloud connection view angle, connecting each connection view angle to obtain a radar point cloud view; drawing a step line map according to the radar point cloud view, exporting the step line map to obtain an open-pit mine present situation plan. The professional mine software can complete the present situation drawing of the mining area based on the UAV aerial survey data; the high-density point cloud data makes the present situation plan more accurate and comprehensive to show the position and elevation of the mining and drainage step slope top, slope bottom and flat plate. For example, Chinese patent CN112634434A discloses a mine three-dimensional model making method based on a UAV, which comprises: obtaining a scene photo of a target area; preprocessing the scene photo to obtain a preprocessed scene photo; dividing the preprocessed scene photo according to regions to obtain multiple groups of preprocessed scene photos and POS data under Gauss 3-degree band projection; and modeling the POS data under Gauss 3-degree band projection to obtain a three-dimensional model. The designed scheme can analyze the mining area coal seam exposure from various angles, and can clearly and intuitively evaluate the mine greening degree and analyze the on-site step distribution and mining and drainage status.

[0005] Although the above design provides a method for measuring the step structure of the open-pit mine using the UAV, the data to be processed is complex and the operation efficiency is low, so there is an urgent need to design a new method for measuring the step structure of the open-pit mine to simplify the data for measuring the step structure of the open-pit mine and improve the operation efficiency. SUMMARY

[0006] In view of the problems of complex data to be processed and low operation efficiency in the existing method for measuring the step structure of the open-pit mine, the present application designs a method for measuring the step structure of the open-pit mine based on UAV image matching point cloud to simplify the data for measuring the step structure of the open-pit mine and improve the operation efficiency.

[0007] A method for measuring the step structure of the open-pit mine based on UAV image matching point cloud, comprising the following steps:

[0008] Step S1, obtaining open-pit mine point cloud data: obtaining open-pit mine image data based on UAV oblique photogrammetry, importing into ContextCapture, and then generating point cloud through image preprocessing, aerial triangulation encryption processing and multi-view image dense matching processing;

[0009] Step S2, open-pit point cloud data preprocessing: based on Kd-Tree to construct neighborhood relationship, based on geometric sampling to realize point cloud simplification, to retain features while simplifying data;

[0010] Step S3, open-pit bench element extraction: according to neighborhood feature extraction slope, select the plane to manually cut the slope, process the slope to get several bench slopes through density clustering, and extract the bench line;

[0011] Step S4, measure the bench structure: according to the extracted bench element to calculate the bench height and bench slope angle.

[0012] An open-pit bench structure measurement method based on unmanned aerial vehicle image matching point cloud, the specific method of step S2 includes:

[0013] Step S21, based on Kd-Tree to construct point cloud neighborhood relationship in three-dimensional space, select the maximum variance dimension as the segmentation axis, set the point with the median value in the segmentation axis direction as the current node, divide the points less than the median value into left subtree nodes, and divide the points greater than the median value into right subtree nodes, and update the segmentation axis to the second largest variance axis and the smallest variance axis in turn, repeat the above steps until all data are processed, at this time, select a point, and the points in the sphere with the point as the center and r as the radius are the neighborhood points of the point, which can be expressed as {p i |||p i -p||<r}, based on Kd-Tree, the nearest neighbor search of data can be quickly completed;

[0014] Step S22, based on geometric sampling to realize point cloud simplification, preset target sampling number U and uniform sampling rate V, calculate the normal angle θ of each point to its neighborhood point, the greater θ is, the greater the curvature of the region is, set the angle threshold η, and the points with θ>η are classified as steep region, then the sampling number of steep region is U*(1-V), and the sampling number of non-steep region is U*V, this method has uniform local area sampling, noise resistance, retains features while reducing data quantity, and improves calculation rate.

[0015] Preferably, the specific method of step S3 includes:

[0016] Step S31, solve the normal vector, calculate the angle θ between the normal vector and the vertical plane, set the threshold value λ, if θ>λ, the point is a slope, otherwise it is a plane;

[0017] Step S32, import the point cloud into cloudcompare, based on straight-through filtering division, so that different benches are separated to the maximum extent;

[0018] Step S33, separate the spatially discrete point clusters by density clustering processing the slope point cloud, to obtain several stepped slopes S={s0, s1, s2,..., s n};

[0019] Step S34, extract the step lines, select a point p i in the stepped slope, there is a neighborhood N(p i ), project the points in N(p i ) onto the neighborhood tangent plane, the projected points form a set Ω, take p i as the corner point, and form a series of included angles Θ={θ1, θ2,..., θ m} in the clockwise direction between the points in Ω and p i , where m=||N(x i )||-1, select the maximum value θ max in Θ, set a threshold value ξ, when θ max >ξ, p i is the boundary point, process all points in the stepped slope to obtain the stepped slope boundary B, there are two step lines and a cross-sectional boundary in B, randomly mark three points in B to fit a plane Π B , calculate the distance from the points in B to Π B , if it is less than the distance threshold ζ, it is an internal point, otherwise it is a noise point, repeat the plane fitting operation until there is no unmarked point in B or the number of iterations is greater than the set maximum number of iterations, at this time the case with the most internal points is the optimal choice, and the corresponding internal point set L B is a step line, remove L B from B, set the points in B to the unmarked state, repeat the plane fitting operation, select the internal point L A corresponding to the most case, and L A is another step line in B.

[0020] Preferably, the specific method of step S4 comprises:

[0021] Step S41, step height calculation, select two adjacent slope lines L A ={a0, a1,..., a n} and L B ={b0, b1,..., b m}, select a point a A in L i , find the nearest point b B of a i in L j , calculate d=||a i -b j ||, d z is the step height at that point, and L ARepeat the above operation at each point to obtain L A The step height set D A ={d1, d2,..., d n} is obtained, and the corresponding L B The corresponding step height set D B ={d'1, d'2,..., d' m} can also be calculated.

[0022] Step S42, step slope angle calculation, the step slope angle at a point is the angle between the normal vector at the point and the vertical surface.

[0023] The beneficial effects obtained by the present application are:

[0024] 1. The open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud designed by the present application can improve the point cloud retrieval speed of open-pit mine step point cloud data based on Kd-Tree neighborhood relationship construction, and realize the simplification of step point cloud based on geometric sampling, while retaining the step characteristics, and improve the technical effect of open-pit mine step structure measurement efficiency.

[0025] 2. The open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud designed by the present application can conveniently obtain image data, and the method is reasonable and the result is reliable, which makes up for the problems of high cost, low efficiency and safety of surveying personnel cannot be guaranteed caused by traditional use of level, total station and GNSS receiver and other machines to measure step structure, so the method has certain application value in open-pit mine step feature extraction, supervision and acceptance and deformation monitoring.

[0026] 3. The open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud designed by the present application can quickly extract open-pit mine step elements and realize structure measurement by processing unmanned aerial vehicle image matching point cloud through geometric characteristics, which can quickly and accurately measure step structure, effectively reduce open-pit mine supervision cost, and improve the usability of unmanned aerial vehicle image matching point cloud in open-pit mine scene.

[0027] The above description is only a summary of the technical scheme of the present application, in order to more clearly understand the technical means of the present application, so as to implement according to the content of the specification, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the following will be described in detail with the preferred embodiments of the present application and the accompanying drawings.

[0028] According to the detailed description of the specific embodiments of the present application in the following text combined with the drawings, those skilled in the art will be more clear about the above and other purposes, advantages and characteristics of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only represent some of the embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative effort. In all the drawings, similar elements or parts are generally identified by similar reference signs. In the drawings, the elements or parts are not necessarily drawn according to the actual proportions.

[0030] Figure 1 A flowchart of an open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud;

[0031] Figure 2 Each step slope surface obtained after density clustering;

[0032] Figure 3 The extracted open-pit mine step line. DETAILED DESCRIPTION

[0033] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. In the following description, specific details such as specific configurations and components are provided only to help a comprehensive understanding of the embodiments of the present application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, in order to be clear and concise, the description of known functions and structures is omitted in the embodiments.

[0034] It should be understood that the "one embodiment" or "the embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "one embodiment" or "the embodiment" appearing throughout the specification does not necessarily mean the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner.

[0035] In addition, the reference numerals and / or letters can be repeated in different examples of the present application. Such repetition is for the purpose of simplification and clarity, and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0036] The term "and / or", used in the present document, only describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" in the present document describes another association relationship of the associated objects, which means that there can be two relationships, for example, A / and B can mean that A exists alone and A and B exist simultaneously. In addition, the character " / " in the present document generally represents an "or" relationship between the associated objects before and after it.

[0037] The term "at least one" in the present document only describes the association relationship of the associated objects, which means that there can be three relationships, for example, at least one of A and B can mean that A exists alone, A and B exist simultaneously, and B exists alone.

[0038] It should also be noted that the relationship terms such as first and second in the present document are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion.

[0039] Embodiment 1

[0040] This embodiment mainly introduces an open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud. For specific flowchart, please refer to Figure 1 , which includes the following steps:

[0041] Step S1, obtaining open-pit mine point cloud data: obtaining open-pit mine image data based on unmanned aerial vehicle oblique photogrammetry, importing into Context Capture, and then generating point cloud through image preprocessing, space triangulation encryption processing and multi-view image dense matching processing;

[0042] Step S2, open-pit mine point cloud data preprocessing: constructing neighborhood relationship based on Kd-Tree, realizing point cloud simplification based on geometric sampling, and simplifying data while retaining features;

[0043] Step S3, open-pit mine step element extraction: extracting slope surface according to neighborhood characteristics, selecting plane to manually divide slope surface, obtaining a plurality of step slope surfaces through density clustering processing of slope surface, and extracting step line;

[0044] Step S4, measuring step structure: calculating step height and step slope angle according to the extracted step element.

[0045] Further, the specific method of step S2 includes:

[0046] Step S21, constructing the point cloud neighborhood relationship in three-dimensional space based on Kd-Tree, selecting the maximum variance dimension as the segmentation axis, setting the point with the median value in the segmentation axis direction as the current node, dividing the points less than the median value into left sub-tree nodes, and dividing the points greater than the median value into right sub-tree nodes, and sequentially updating the segmentation axis to the second largest variance axis and the smallest variance axis, and repeating the above steps until all data are processed, at this time, a point is selected, and the points in the sphere with the selected point as the center and r as the radius are the neighborhood points of the selected point, which can be expressed as {p i |||p i -p||<r} and the nearest neighbor search of data can be quickly completed based on Kd-Tree;

[0047] Step S22, realizing point cloud simplification based on geometric sampling, presetting a target sampling number U and a uniform sampling rate V, calculating the normal angle θ of each point to its neighborhood points, the greater θ is, the greater the curvature of the region is, setting an angle threshold η, and classifying the points with θ>η into steep regions, so that the sampling number of the steep regions is U*(1-V) and the sampling number of the non-steep regions is U*V, the method has local region sampling uniformity and noise resistance, can reduce the data amount while retaining features, and improves the calculation rate.

[0048] Further, the specific method of step S3 comprises:

[0049] Step S31, solving the normal vector, calculating the angle θ between the normal vector and the vertical plane, setting a threshold λ, and if θ>λ, the point is a slope surface, otherwise, the point is a plane;

[0050] Step S32, importing the point cloud into cloudcompare, and cutting based on the straight-through filter to separate different steps to the greatest extent;

[0051] Step S33, processing the slope surface point cloud through density clustering to separate the discrete point clusters in space and obtain a plurality of step slope surfaces S={s0,s1,s2,...,s n};

[0052] Step S34, extracting the step line, selecting a point p i in the step slope surface, and existing a neighborhood N(p i ), projecting the points in N(p i ) onto the neighborhood tangent plane, and constructing a set Ω with the projected points, taking p i as a corner point, forming a series of angles Θ={θ1,θ2,...,θ m} between the points in Ω and p i in a clockwise direction, wherein m=||N(x i )||-1, selecting the maximum value θ max in Θ, and setting a threshold value ξ, when θ max >ξ, pi That is the boundary point, processing all points in the step slope, get step slope boundary B, B there are two steps and cross-sectional boundary line, in B randomly marked 3 points to fit the plane Π B , calculate the distance of B points to Π B If less than the distance threshold ζ is an internal point, the opposite is a noise point, repeat the plane fitting operation until B can not find the unmarked points or the number of iterations is greater than the maximum number of iterations set, at this time the matching most internal points is the optimal choice, corresponding to the internal point set L B That is a step line, L B In B, the points are set to unmarked state, repeat the plane fitting operation, select the most corresponding internal points L A , L A Is another step line in B.

[0053] Further, the specific method of step S4 comprises:

[0054] Step S41, step height calculation, select L A ={a0, a1,...,a n} and L B ={b0, b1,...,b m} two adjacent slope line, select a A point in L i , find the nearest point b B of a i in L j , calculate d = ||a i -b j ||, d z is the step height at that point, repeat the above operation for each point in L A to get the step height set D A ={d1, d2,...,d n} of L B , the corresponding L B can also be calculated to get the corresponding step height set D m ={d'1, d'2,...,d' i};

[0055] Step S42, step slope angle calculation, the step slope angle at a point is the angle between the normal vector of the point and the vertical surface.

[0056] The open-pit mine step structure measurement method based on the unmanned aerial vehicle image matching point cloud is designed, the neighborhood relationship is constructed based on Kd-Tree, the point cloud retrieval speed of the open-pit mine step point cloud data can be improved, the step point cloud is simplified based on geometric sampling, the technical effects of simplifying the data while reserving the step characteristics and improving the open-pit mine step structure measurement efficiency are achieved.

[0057] Embodiment 2

[0058] Based on the above-mentioned embodiment 1, the embodiment mainly introduces the optimization of the open-pit mine step structure measurement method based on the unmanned aerial vehicle image matching point cloud, including the following steps.

[0059] Step S1, data acquisition, the open-pit mine image data is acquired based on the unmanned aerial vehicle oblique photogrammetry, is imported into ContextCapture, and then the point cloud is generated through the steps of image preprocessing, aerial triangulation encryption and multi-view image dense matching.

[0060] Step S2, point cloud data preprocessing, the neighborhood relationship is constructed based on Kd-Tree, the point cloud is simplified based on geometric sampling, the data is simplified while the characteristics are reserved, and the process is as follows:

[0061] Step S21, the point cloud neighborhood relationship is constructed in the three-dimensional space based on Kd-Tree, the maximum variance dimension is selected as the segmentation axis, the point with the median value in the segmentation axis direction is set as the current node, the points less than the median value are divided into left subtree nodes, and the points greater than the median value are right subtree nodes, the segmentation axis is updated to the second largest variance axis and the smallest variance axis in turn, and the above steps are repeated until all data are processed, at this time, a point is selected, the points in the ball with the center of the point and the radius r are the neighborhood points of the point, which can be expressed as {p i |||p i -p||<r} to quickly complete the nearest neighbor search of data;

[0062] Step S22, the point cloud is simplified based on geometric sampling, the target sampling number is 1 million and the uniform sampling rate is 0.3, the normal angle θ of each point to its neighborhood point is calculated, the greater the angle θ, the greater the curvature of the region, the angle threshold η is set, the points with θ>η are classified as steep regions, the sampling number of the steep region is 700,000, and the sampling number of the non-steep region is 300,000, the method uniformly samples the local region, has noise resistance, reduces the data amount while reserving the characteristics, and improves the calculation rate.

[0063] Step S3, open-pit mine step element extraction, the slope surface is extracted according to the neighborhood characteristics, the slope surface is divided based on the straight-through filter, 17 step slope surfaces are obtained by processing the slope surface through density clustering, and the step line is extracted, and the process is as follows:

[0064] Step S31: Solve for the normal vector and calculate the angle θ between the normal vector and the vertical plane. Set a threshold of 16°. If θ > 16°, the point is a slope; otherwise, it is a plane.

[0065] Step S32: Import the point cloud into CloudCompare, select the planes y=520 and x=420 to manually divide the slope, so that the different steps are separated to the greatest extent.

[0066] Step S33: By processing the slope point cloud through density clustering, the spatially discrete point clusters are separated to obtain several stepped slope surfaces S = {s0, s1, s2, ..., s}. 16};

[0067] Figure 2 These are the slope surfaces of each step obtained after density clustering.

[0068] Step S34: Extract the step line and select a point p on the slope of the step. i There exists a neighborhood N(p) i ), will N(p i Points in a region are projected onto a neighborhood tangent plane, and the projected points form a set Ω, with p as the reference point. i As corner points, connect each point in Ω to p in a clockwise direction. i Forming a series of included angles Θ = {θ1, θ2, ..., θ m}, where m=||N(x i )||-1, select the maximum value θ among Θ max Set a threshold ξ, when θ max When p > ξ i These are the boundary points. Process all points on the stepped slope to obtain the stepped slope boundary B. B contains two step lines and a cross-sectional boundary. Randomly mark three points in B to fit a plane Π. B Calculate the distance from the midpoint of B to Π. B If the distance to the unmarked point is less than the distance threshold ζ, it is considered an interior point; otherwise, it is considered a noise point. The plane fitting operation is repeated until no unmarked points are found in B or the number of iterations exceeds the set maximum number of iterations. The optimal choice is the one that matches the most interior points, corresponding to the interior point set L. B That is, a step line, in B, remove L. B Set the midpoint of B to the unlabeled state, repeat the plane fitting operation, and select the interior point L corresponding to the best case. A L A This is another step line in B.

[0069] Figure 3 This refers to the extracted open-pit mine bench line.

[0070] Step S4, step structure measurement, calculate the step height and step slope angle according to the extracted step element, the process is as follows:

[0071] Step S41, step height calculation, select L A ={a0, a1,...,a n} and L B ={b0, b1,...,b m} two adjacent slope lines, select L A point a i , find the nearest point b B of a i in L j , calculate d = ||a i -b j ||, d z is the step height at this point, repeat the above operation for each point in L A to get the step height set D A ={d1, d2,...,d n} of L B , the corresponding L B can also be calculated to get the corresponding step height set D m ={d'1, d'2,...,d' };

[0072] Step S42, step slope angle calculation, the step slope angle at a point is the angle between the normal vector at the point and the vertical surface.

[0073] The step structure measurement method based on unmanned aerial vehicle image matching point cloud designed in the application can conveniently obtain image data, and the method is reasonable and the result is reliable, which makes up for the problems of high cost, low efficiency and safety of surveying personnel cannot be guaranteed caused by traditional use of level, total station and GNSS receiver and other machines to measure step structure, so the method has certain application value in step feature extraction, supervision and acceptance and deformation monitoring of open-pit mine.

[0074] The step structure measurement method based on unmanned aerial vehicle image matching point cloud designed in the application can quickly extract open-pit mine step elements and realize structure measurement by geometric feature processing of unmanned aerial vehicle image matching point cloud, which can quickly and accurately measure step structure, effectively reduce open-pit mine supervision cost, and improve the usability of unmanned aerial vehicle image matching point cloud in open-pit mine scene.

[0075] The above merely describes the preferred embodiments of the present application, and is not intended to limit the protection scope of the present application. The present application can have various changes and modifications for those skilled in the art. Any changes, modifications, replacements, integrations and parameter changes to the embodiments within the spirit and principle of the present application, which can realize the same functions without departing from the principle and spirit of the present application, fall within the protection scope of the present application.

Claims

1. An open-pit mine step structure measurement method based on unmanned aerial vehicle image matching point cloud, characterized in that, The method comprises the following steps: Step S1, obtaining open-pit point cloud data: obtaining open-pit image data based on unmanned aerial oblique photogrammetry, importing into Context Capture, and then generating point cloud through image preprocessing, aerial triangulation encryption processing and multi-view image dense matching processing; Step S2, open-pit point cloud data preprocessing: constructing neighborhood relationship based on Kd-Tree, and realizing point cloud simplification based on geometric sampling, so as to retain features while simplifying data; Step S3, open-pit step element extraction: extracting slope surface according to neighborhood features, selecting plane to manually cut slope surface, obtaining a plurality of step slope surfaces through density clustering processing of the slope surface, and extracting step lines; Step S4, measuring step structure: calculating step height and step slope angle according to the extracted step element, and the step slope angle at a point being the included angle between the normal vector of the point and the vertical surface; The specific method of step S2 comprises: Step S21, constructing point cloud neighborhood relationship in three-dimensional space based on Kd-Tree, selecting the maximum variance dimension as the segmentation axis, setting the point with the median value in the segmentation axis direction as the current node, dividing the points less than the median value into left sub-tree nodes, and dividing the points greater than the median value into right sub-tree nodes, sequentially updating the segmentation axis to the second largest variance axis and the smallest variance axis, repeating the above steps until all data are processed, at this time, a point is selected, and the points in the ball with the center of the point and the radius of are the neighborhood points of the point, which can be expressed as , and the nearest neighbor search of data can be quickly completed based on Kd-Tree; Step S22, realizing point cloud simplification based on geometric sampling, and presetting a target sampling number and a uniform sampling rate V, calculating the normal angle of each point to its neighborhood points , The greater the neighborhood curvature of the point is, and a preset angle threshold is set Points of are classified into steep regions, and the sampling number of the steep regions is , and the sampling number of the non-steep regions is The method for realizing point cloud simplification based on geometric sampling has uniform local region sampling, anti-noise, reduces the data amount while preserving the features, and improves the calculation rate. The specific method of step S3 comprises: Step S31, solving the normal vector, calculating the angle between the normal vector and the vertical plane , setting a threshold value , if then the point is a slope surface, otherwise it is a plane; Step S32, importing the point cloud into cloud compare, and cutting based on straight-through filtering, so that different steps are separated to the maximum extent; Step S33, separate the spatially discrete point clusters by density clustering processing the slope point cloud to obtain several stepped slopes ; Step S34, extracting step lines, selecting a point in step slope , there is a neighborhood , the point in is projected onto the neighborhood tangent plane, and the projected point constitutes a set , taking as the corner point, the points in are arranged in a clockwise direction two by two with to form a series of included angles , wherein , the maximum value in is selected , the threshold value is set , when , is the boundary point, all points in the step slope are processed to obtain the step slope boundary , there are two step lines and cross-sectional boundaries in B, three points in are randomly marked to fit a plane , the distance from the point in to is calculated, if it is less than the distance threshold , it is an internal point, otherwise it is a noise point, repeat the plane fitting operation until no unmarked point is found in or the number of iterations is greater than the set maximum number of iterations, at this time the case with the most internal points is the optimal choice, and the corresponding internal point set is a step line, remove from , set the points in B to unmarked state, repeat the plane fitting operation, and select the internal points corresponding to the most case , is another step line in B.

2. The method of claim 1, wherein: The specific method of step S4 comprises: Step S41, step height calculation, selected and two adjacent slope lines, select middle point , in find the nearest point , calculate , step height at the point, for each point in the above operation to get step height set , the corresponding can also be calculated to the corresponding step height set ; Step S42, step slope angle calculation.

3. The method of claim 1, wherein the method further comprises: In the step S22, the preset target sampling number The preset target sampling number is 1 million, the uniform sampling rate V is 0.3, the sampling number of the steep region is 700,000, and the sampling number of the non-steep region is 300,000.

4. The method of claim 1, wherein, In step S31, the threshold is set to 16°, and if then the point is a slope, otherwise it is a flat surface.

5. A method for measuring bench structure of an open-pit mine based on matching point clouds of UAV images according to any one of claims 1 or 4, characterized in that, In step S32, the method based on the straight-through filtering segmentation is: selecting a plane and Manually segment the slope.

Images