Power transmission conductor three-dimensional feature processing method and system based on micro-terrain point cloud space and storage medium

Abstract

The application provides a kind of transmission conductor three-dimensional feature processing method, system and storage medium based on microtopography point cloud space, elevation histogram adaptive ground filtering processing is carried out to the original point cloud data set of transmission corridor, the point cloud frequency distribution of each elevation interval is counted to adaptively determine the segmentation threshold, remove ground point cloud to generate overhead point cloud set;Overhead point cloud set is voxelized spatial connectivity clustering processing, and it is clustered into tower component and conductor component;The axial feature extraction of principal component analysis is carried out to each independent connected component, and the minimum eigenvalue corresponding eigenvector is used as the main extension direction, and the component with the included angle less than the preset acute angle threshold with horizontal plane is marked as conductor component;Conductor component is carried out three times spline interpolation spatial curve fitting, and discrete point cloud is fitted into conductor catenary parameter equation, and the three-dimensional space trajectory parameter of conductor is output.The application significantly improves the automation degree of transmission conductor three-dimensional feature extraction.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection of transmission lines and three-dimensional point cloud data processing technology. Specifically, it relates to a method, system and storage medium for three-dimensional feature processing of transmission conductors based on micro-topographic point cloud space. Background Technology

[0002] With the rapid development of UAV-borne LiDAR technology, using point cloud data for refined 3D modeling of power transmission corridors has become an important method for line inspection. However, power transmission corridor scenarios are highly complex. The raw point cloud data often contains multiple targets simultaneously, including ground vegetation, buildings, tower structures, and overhead conductors, and these targets exhibit significant differences in point cloud density and geometric overlap in 3D space. Existing technologies typically use fixed-parameter ground filtering algorithms or clustering methods based on single geometric features for target segmentation when processing such data. When faced with power transmission corridors under complex micro-topographical conditions, these methods often struggle to adaptively determine the boundary between ground and non-ground areas, leading to the mixing of tower point clouds and conductor point clouds or the accidental deletion of valid point clouds.

[0003] Furthermore, existing technologies, when distinguishing between towers and conductors, largely rely on manually set geometric rules or simple linear feature judgments, failing to fully utilize the axial distribution characteristics of point cloud components in space for automated identification. This easily leads to misclassifying tilted tower components as conductors, or conductors with large sag as tower attachments. In addition, when acquiring the three-dimensional spatial trajectory of conductors, existing technologies typically directly perform linear fitting or simple polynomial fitting on discrete point clouds, ignoring the catenary physical morphology naturally formed by the conductor under gravity. This results in a significant deviation between the fitting results and the actual spatial geometry of the conductor, failing to provide high-precision parametric representations for subsequent applications such as line sag analysis and crossover distance calculation. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for three-dimensional feature processing of power transmission lines based on micro-topographic point cloud space, the method comprising: Load the original point cloud data set of the power transmission corridor collected by the UAV airborne lidar. The original point cloud data set of the power transmission corridor includes the three-dimensional spatial coordinate information and echo intensity reflectivity information of each scanning point. The original point cloud data set of the power transmission corridor is subjected to elevation histogram adaptive ground filtering processing. The point cloud frequency distribution in each elevation interval is statistically analyzed. The segmentation threshold between ground points and non-ground points is determined. Ground point clouds are removed to generate an overhead point cloud set, which includes tower point clouds and conductor point clouds. Voxelized spatial connectivity clustering is performed on the overhead point cloud set. The point cloud space is divided into a regular voxel grid with a preset voxel size. The continuity of point cloud density between adjacent voxels is determined. The overhead point cloud set is clustered into multiple independent connected components. The independent connected components include tower components and conductor components. Principal component analysis is performed on each of the independent connected components to extract axial features. The feature vector corresponding to the minimum feature value of the point cloud of each component is calculated as the main extension direction. Components whose main extension direction has an angle with the horizontal plane that is less than a preset acute angle threshold are marked as wire components. The conductor assembly is subjected to cubic spline interpolation spatial curve fitting processing to fit the discrete point cloud into a smooth catenary parametric equation, and the three-dimensional spatial trajectory parameters of the conductor are output.

[0005] Furthermore, embodiments of the present invention also provide a three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space, comprising: A processor; a machine-readable storage medium for storing machine-executable instructions of the processor; wherein the processor is configured to execute the above-described method for processing three-dimensional features of power transmission lines based on micro-topographic point cloud space by executing the machine-executable instructions.

[0006] In another aspect, embodiments of the present invention also provide a computer program product, the computer program product including machine-executable instructions, the machine-executable instructions being stored in a computer-readable storage medium, the processor of the power transmission line three-dimensional feature processing system based on micro-topographic point cloud space reading the machine-executable instructions from the computer-readable storage medium, the processor executing the machine-executable instructions, causing the power transmission line three-dimensional feature processing system based on micro-topographic point cloud space to execute the above-described power transmission line three-dimensional feature processing method based on micro-topographic point cloud space.

[0007] Based on the above, by performing elevation histogram adaptive ground filtering on the raw point cloud data set of the power transmission corridor collected by UAV airborne lidar, the segmentation threshold between ground points and non-ground points can be adaptively determined according to the statistical characteristics of the point cloud frequency distribution in each elevation interval. This effectively addresses the problem of uneven ground elevation under complex micro-topography conditions, accurately removing ground point clouds while preserving the overhead target point clouds to the greatest extent. On this basis, the overhead point cloud set is divided into regular voxel grids by voxelized spatial connectivity clustering and the continuity of point cloud density between adjacent voxels is judged, realizing the natural spatial separation of tower components and conductor components, and avoiding the target aliasing problem caused by point cloud density differences in traditional methods. Furthermore, by performing principal component analysis to extract axial features from each independent connected component, the eigenvector corresponding to the minimum eigenvalue is used as the main extension direction. Components whose main extension direction forms an angle less than a preset acute angle threshold with the horizontal plane are marked as conductor components. This fully utilizes the essential geometric difference between the near-horizontal extension of conductors and the near-vertical extension of towers in space, achieving high-precision automated differentiation between towers and conductors, overcoming the misjudgment problem caused by relying on simple geometric rules in existing technologies. Finally, by performing cubic spline interpolation spatial curve fitting on the conductor components, the discrete point cloud is fitted into parametric equations that conform to the physical morphology of the catenary conductor, outputting the three-dimensional spatial trajectory parameters of the conductor. This ensures that the fitting result is not only geometrically smooth and continuous but also physically reflects the natural spatial morphology of the conductor under gravity and tension. Therefore, the automation level, recognition accuracy, and usability of the three-dimensional feature extraction of transmission conductors are significantly improved. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the execution flow of the three-dimensional feature processing method for power transmission lines based on micro-topographic point cloud space provided in an embodiment of the present invention.

[0009] Figure 2 This is a logical schematic diagram of the three-dimensional feature processing method for power transmission lines based on micro-topographic point cloud space provided in the embodiments of the present invention.

[0010] Figure 3 This is one of the simulation diagrams of the three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space provided in the embodiments of the present invention.

[0011] Figure 4 This is the second simulation diagram of the three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space provided in this embodiment of the invention.

[0012] Figure 5 This is the third simulation diagram of the three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space provided in this embodiment of the invention. Detailed Implementation

[0013] Figure 1 This is a flowchart illustrating a method for processing three-dimensional features of power transmission lines based on micro-topographic point cloud space, provided in one embodiment of the present invention. The following is a detailed explanation in conjunction with... Figure 2 A detailed introduction is provided. This embodiment uses an overhead power transmission line corridor traversing hilly terrain as the processing object, and performs 3D feature processing on the UAV LiDAR point cloud data of this corridor. The data acquisition has obtained airspace approval and geographic information authorization, the point cloud does not contain personal privacy, and complies with legal and regulatory requirements.

[0014] Step S110: Load the original point cloud data set of the power transmission corridor collected by the UAV airborne lidar. The original point cloud data set of the power transmission corridor includes the three-dimensional spatial coordinate information and echo intensity reflectivity information of each scanning point.

[0015] A drone equipped with a lidar system flies along a planned route along the power transmission corridor. The lidar system emits laser beams to the ground and objects using a pulsed scanning method and receives echo signals. Through combined navigation calculations using an onboard global navigation satellite system and an inertial measurement unit, and combining laser ranging values ​​and scanning angle values, the three-dimensional spatial coordinates of each laser footpoint in the agreed-upon Earth coordinate system are calculated. These three-dimensional spatial coordinates include the X-axis coordinate component along the east direction, the Y-axis coordinate component along the north direction, and the Z-axis coordinate component along the elevation direction of the ellipsoid normal. Simultaneously, the lidar receiving system records the ratio of the peak power of the echo pulse to the peak power of the emitted pulse at each scanning point, and after system calibration and conversion, obtains the echo intensity reflectivity I. The original point cloud data set of the power transmission corridor is denoted as P0, containing N0 scanning points. The data record of the k-th scanning point is (Xk, Yk, Zk, Ik), k=1, 2, ..., N0. P0 is stored in the memory of the data processing device in LAS standard format.

[0016] Step S120: Perform elevation histogram adaptive ground filtering on the original point cloud data set of the power transmission corridor, statistically analyze the point cloud frequency distribution in each elevation interval, determine the segmentation threshold between ground points and non-ground points, remove ground point clouds, and generate an overhead point cloud set, which includes tower point clouds and conductor point clouds.

[0017] Step S121: Extract the elevation direction coordinate components of each scanning point from the original point cloud data set of the power transmission corridor, generate an elevation value sequence of all scanning points, search for the maximum and minimum elevation values ​​in the elevation value sequence of all scanning points, and construct the total elevation distribution interval.

[0018] Extract the elevation direction coordinate components Zk of each scan point from P0 point by point to form a one-dimensional sequence Z, Z=[Z1, Z2, ..., ZN0]. Traverse all N0 elements in Z and find the maximum elevation value Amax and the minimum elevation value Amin by pairwise comparison. The total elevation distribution interval is determined to be [Amin, Amax].

[0019] Step S122: Divide the total elevation distribution interval evenly into multiple elevation statistical intervals, traverse the elevation value sequence of all scan points, assign each scan point to the corresponding elevation statistical interval according to its elevation value, count the cumulative frequency of scan points falling into each elevation statistical interval, and generate an elevation interval point cloud frequency distribution sequence.

[0020] Let K be the number of divisions, where K is a positive integer. Divide [Amin, Amax] evenly into K contiguous subintervals, each with a width of W = (Amax - Amin) / K. Iterate through each Zk in Z, j = ceil((Zk - Amin) / W), and increment the cumulative frequency fj of the j-th subinterval by 1, where j = 1, 2, ..., K. After the iteration is complete, the frequency distribution sequence F = [f1, f2, ..., fK] is obtained.

[0021] Step S123: Perform frequency difference operation on the frequency distribution sequence of the point cloud in the elevation interval to obtain the frequency change gradient between each adjacent elevation statistical interval. Determine the boundary of the interval with the largest absolute value of the frequency change gradient as the initial segmentation elevation threshold between the ground point and the overhead point. Construct a local elevation neighborhood window with the initial segmentation elevation threshold as the center. Extract the scan points in the local elevation neighborhood window to form a neighborhood point set. Calculate the elevation numerical standard deviation of the neighborhood point set and compare the elevation numerical standard deviation with the preset elevation fluctuation tolerance value.

[0022] Perform a first-order forward difference on F, gj = fj + 1 - fj, j = 1, 2, ..., K-1. Calculate each |gj| and find the index m that maximizes |gj|. Initially select the segmentation elevation threshold C = Amin + m × W. Construct a neighborhood window [CD, C+D] centered on C, where D is the preset neighborhood half-width. Extract all scan points Zk ∈ [CD, C+D] from P0 to form a neighborhood point set Q, containing T points. Calculate the elevation standard deviation σ of Q = sqrt(Σ(Zu-μ)² / (T-1)), where μ is the arithmetic mean of all elevation values ​​in Q, summed over all points in Q. Compare σ with the preset elevation fluctuation tolerance value σth.

[0023] Step S124: When the standard deviation of the elevation value is less than the preset elevation fluctuation tolerance value, the initial selected segmentation elevation threshold is determined as the final segmentation elevation threshold. The original point cloud data set of the power transmission corridor is binary segmented using the final segmentation elevation threshold. Scanning points with elevation values ​​less than the final segmentation elevation threshold are classified as ground point clouds, and scanning points with elevation values ​​greater than or equal to the final segmentation elevation threshold are classified as overhead point clouds.

[0024] When σ < σth, C is determined as the final segmentation elevation threshold Zth. For each scan point in P0, if Zk < Zth, it is marked as a ground point cloud; if Zk ≥ Zth, it is marked as an overhead point cloud.

[0025] Step S125: Remove all scan points classified as ground point clouds from the original point cloud data set of the power transmission corridor, and retain all scan points classified as overhead point clouds to form an overhead point cloud set. The overhead point cloud set includes the three-dimensional spatial coordinate information and echo intensity reflectivity information of each overhead scan point.

[0026] Remove all ground point cloud marker scan points from P0 and retain all overhead point cloud marker scan points to form an overhead point cloud set P1. P1 contains N1 scan points, and the data format of each scan point is (Xk, Yk, Zk, Ik).

[0027] Step S126: Perform point cloud spatial distribution statistics on the overhead point cloud set, divide the overhead point cloud set into multiple elevation layers according to the elevation direction, count the number of overhead scanning points in each elevation layer, generate an overhead point cloud elevation layer distribution histogram, and extract the clustered and sparse segments of the point cloud in the elevation direction based on the changing trend of the number of points in each elevation layer in the overhead point cloud elevation layer distribution histogram. Pre-mark the scanning points corresponding to the clustered segments as potential point clouds of towers, and pre-mark the scanning points corresponding to the sparse segments as potential point clouds of conductors.

[0028] Extract the maximum value Hmax and minimum value Hmin of the Z-coordinate component from P1. Divide the elevation range [Hmin, Hmax] into L elevation layers uniformly. Count the number of scan points falling into each elevation layer to generate a histogram H1 of the elevation layer distribution of the overhead point cloud. Apply a one-dimensional mean smoothing filter to H1. Identify local peak and valley intervals of the point cloud count on the smoothed H1. Pre-label the scan points corresponding to the local peak intervals as potential point clouds for towers, and pre-label the scan points corresponding to the local valley intervals as potential point clouds for conductors.

[0029] Step S127: Perform local point cloud density calculations on the potential point cloud of the tower and the potential point cloud of the conductor respectively to obtain the point cloud density values ​​of the potential area of ​​the tower and the potential area of ​​the conductor. Based on the density difference characteristics between the point cloud density values ​​of the potential area of ​​the tower and the potential area of ​​the conductor, generate the preliminary separation marking results of the tower and conductor.

[0030] For the set of scan points pre-labeled as potential point clouds for towers, the local point cloud density of each scan point is calculated using a fixed-radius sphere neighborhood search method. Centered on scan point p, the number of neighboring scan points contained within a sphere of radius R is counted, and this number is divided by the volume of the sphere to obtain the local point cloud density value of point p. The local point cloud density is calculated in the same way for the set of scan points pre-labeled as potential point clouds for conductors. The average density value Dt of all points in the potential tower region and the average density value Dw of all points in the potential conductor region are calculated. For scan points located in the boundary region, if the difference between the local density value of that point and Dt is less than the difference between Dt and Dw, the point is corrected to be a potential point cloud for towers; otherwise, it is corrected to be a potential point cloud for conductors. Finally, the preliminary separation and labeling results for towers and conductors are generated.

[0031] Step S130: Perform voxelized spatial connectivity clustering processing on the overhead point cloud set, divide the point cloud space into regular voxel grids with a preset voxel size, determine the continuity of point cloud density between adjacent voxels, and cluster the overhead point cloud set into multiple independent connected components, the multiple independent connected components including tower components and conductor components.

[0032] Step S131: Extract the three-dimensional spatial coordinate information of all overhead scanning points from the overhead point cloud set, search for the minimum and maximum coordinate values ​​along the three-dimensional coordinate axis direction, construct a three-dimensional spatial bounding box that completely surrounds the overhead point cloud set, and record the span range of the three-dimensional spatial bounding box in each coordinate axis direction.

[0033] Extract the 3D coordinates of all N1 scan points from P1. Search for the minimum and maximum coordinate values ​​Xmin and Xmax along the X-axis, Ymin and Ymax along the Y-axis, and Zmin and Zmax along the Z-axis. Construct a 3D bounding box with diagonal vertices (Xmin, Ymin, Zmin) and (Xmax, Ymax, Zmax). The spans of the three coordinate axes are Lx = Xmax - Xmin, Ly = Ymax - Ymin, and Lz = Zmax - Zmin, respectively.

[0034] Step S132: Using the preset voxel side length as the basic dividing unit, the three-dimensional bounding box is uniformly divided along the three coordinate axes to generate a voxel grid network with row index number, column index number and layer index number. Each voxel grid unit in the voxel grid network corresponds to a cubic subspace region in the three-dimensional space.

[0035] Let the voxel side length be V, in meters. Divide the bounding box uniformly along the X-axis into Nx = ceil(Lx / V) rows, along the Y-axis into Ny = ceil(Ly / V) columns, and along the Z-axis into Nz = ceil(Lz / V) layers. The voxel grid network contains Nx × Ny × Nz voxel grid cells. The voxel grid cell in the i-th row, j-th column, and k-th layer corresponds to the 3D spatial range [Xmin + (i-1)V, Xmin + iV], [Ymin + (j-1)V, Ymin + jV], [Zmin + (k-1)V, Zmin + kV], i = 1, 2, ..., Nx, j = 1, 2, ..., Ny, k = 1, 2, ..., Nz.

[0036] Step S133: Traverse each overhead scanning point in the overhead point cloud set, calculate the row index number, column index number and layer index number to which the overhead scanning point belongs based on the three-dimensional spatial coordinate information of the overhead scanning point, and assign the overhead scanning point to the corresponding voxel grid cell to complete the voxel mapping of all overhead scanning points.

[0037] Iterate through each scan point p in P1, whose 3D coordinates are (Xp, Yp, Zp). The row index i = ceil((Xp-Xmin) / V), the column index j = ceil((Yp-Ymin) / V), and the layer index k = ceil((Zp-Zmin) / V). Store the index information of scan point p into the point cloud list of the voxel raster cell (i, j, k).

[0038] Step S134: Count the number of overhead scan points contained in each voxel grid cell in the voxel grid network, mark voxel grid cells with a number of overhead scan points greater than zero as occupied voxels, mark voxel grid cells with a number of overhead scan points equal to zero as vacant voxels, and generate a binary labeling map of voxel occupancy status.

[0039] Traverse all voxel grid cells in the voxel grid network and count the length of the point cloud list within each voxel grid cell. If the length is greater than 0, mark the voxel grid cell as 1; if the length is equal to 0, mark it as 0. Generate a three-dimensional binary array O of size Nx×Ny×Nz, where O(i, j, k) takes the value 1 or 0.

[0040] Step S135: Perform a neighborhood connectivity scan on the binary marker map of voxel occupancy status, traverse each voxel grid cell marked as an occupancy voxel, check the occupancy status of the six adjacent voxel grid cells of the voxel grid cell in the face adjacent direction, and record the adjacent voxel grid cells that are adjacent to the occupancy voxel and are also occupancy voxels as connected adjacent voxels.

[0041] For each voxel grid cell (i, j, k) in O with a value of 1, define six adjacent directions as (i+1, j, k), (i-1, j, k), (i, j+1, k), (i, j-1, k), (i, j, k+1), and (i, j, k-1). Check the values ​​of these six adjacent voxel grid cells in O one by one. If an adjacent voxel grid cell has a value of 1 in O, then record that adjacent voxel grid cell as a connected adjacent voxel of the current voxel grid cell.

[0042] Step S136: Using each occupied voxel that is not assigned to any connected component as a seed voxel, perform a three-dimensional spatial region growth operation, merge the seed voxel and all connected neighboring voxels directly connected to the seed voxel into a connected voxel cluster, and use the newly merged occupied voxels into the connected voxel cluster as new growth front voxels to continue searching for their respective connected neighboring voxels.

[0043] Initialize an access marker array of the same size as O, with all elements initialized to 0. Initialize the connected component list to empty. Iterate through all voxel raster cells in O with a value of 1. If the access marker of a cell is 0, set that cell as the seed voxel, create a new empty queue, add the seed voxel to the queue, and set its access marker to 1. Execute a 3D spatial region growth loop: remove the head voxel raster cell from the queue, search for its six adjacent voxel raster cells in adjacent directions. If an adjacent voxel raster cell in O has a value of 1 and an access marker of 0, add that adjacent voxel raster cell to the queue and set its access marker to 1. When the queue is empty, end the loop, group all voxel raster cells visited in this loop into a connected voxel cluster, and store it in the connected component list.

[0044] Step S137: When the voxel at the growth front of the connected voxel cluster can no longer find an unvisited connected neighboring voxel, the growth process of the connected voxel cluster is terminated. The set of overhead scanning points contained in all occupied voxels within the connected voxel cluster is extracted as an independent connected component. Component scale feature analysis is performed on the independent connected component, and the projected span length of the independent connected component in each coordinate axis direction in three-dimensional space is calculated. The spatial morphological extension of the independent connected component is determined according to the ratio of the maximum span direction to the minimum span direction in the projected span length. Independent connected components with spatial morphological extension greater than a preset extension threshold are initially classified as conductor components, and independent connected components with spatial morphological extension less than or equal to the preset extension threshold are initially classified as tower components, generating preliminary classification results for independent connected components.

[0045] For each connected voxel cluster in the connected component list, merge the point cloud lists of all voxel raster cells within it, and extract all scan points contained in that independent connected component. Calculate the projected span of the component: the projected span in the X-axis direction is the maximum value minus the minimum value of the X-coordinate of the point cloud within the component; similarly, calculate the projected spans in the Y-axis and Z-axis directions. Take the maximum and minimum values ​​of the projected spans in the three directions; the spatial morphological extension is equal to the maximum value divided by the minimum value. If the extension is greater than a preset extension threshold, it is initially classified as a conductor component; if the extension is less than or equal to the preset extension threshold, it is initially classified as a tower component.

[0046] Step S138: Perform cross-component spatial adjacency analysis on the preliminary classification results of the independent connected components, extract the number of adjacency faces between the independent connected components initially classified as conductor components and the independent connected components initially classified as tower components in the voxel grid network, and correct the classification label of the independent connected components near the boundary based on the number of adjacency faces.

[0047] For independent connected components initially classified as conductor components and independent connected components initially classified as pole components, the number of adjacent mating surfaces between the boundary voxel grid cells of the conductor component and the boundary voxel grid cells of the pole component is searched in the voxel grid network. If the number of mating surfaces is greater than a preset threshold for the number of adjacent surfaces, and the difference between the spatial morphological extension of the conductor component and a preset extension threshold is within a preset range, then the classification label of the conductor component is corrected to pole component.

[0048] Step S140: Perform principal component analysis on each of the independent connected components to extract axial features, calculate the feature vector corresponding to the minimum feature value of the point cloud of each component as the main extension direction, and mark the components whose main extension direction has an angle with the horizontal plane less than a preset acute angle threshold as wire components.

[0049] Step S141: For each independent connected component, extract the three-dimensional spatial coordinate information of all scan points within the independent connected component, construct a local coordinate system with the centroid of the independent connected component as the origin, transform the three-dimensional spatial coordinate information of all scan points into the local coordinate system to obtain a centroid-free spatial coordinate set, and use the centroid-free spatial coordinate set to construct the point cloud spatial scatter matrix of the independent connected component. Each element of the point cloud spatial scatter matrix is ​​the coordinate component covariance of all scan points within the independent connected component in the corresponding coordinate axis direction.

[0050] For each independent connected component, extract the 3D coordinates of all M scan points within it. Calculate the centroid coordinates: the X-axis centroid is the sum of the X-coordinates of all points divided by M; the Y-axis and Z-axis centroids are calculated similarly. Calculate the centroid-free coordinates for each scan point, i.e., the point coordinates minus the centroid coordinates. The point cloud spatial scatter matrix is ​​a 3x3 symmetric matrix. The element in the first row and first column of the matrix is ​​the sum of the squares of the centroid-free X-coordinates of all points divided by M. The element in the second row and second column is the sum of the squares of the centroid-free Y-coordinates of all points divided by M. The element in the third row and third column is the sum of the squares of the centroid-free Z-coordinates of all points divided by M. The elements in the first row and second column, and the second row and first column are both the sum of the products of the centroid-free X-coordinates and centroid-free Y-coordinates of all points divided by M. The elements in the first row and third column, and the third row and first column are both the sum of the products of the centroid-free X-coordinates and centroid-free Z-coordinates of all points divided by M. The elements in the second row and third column, and the third row and second column, are both the sum of the products of the centroid-free Y-coordinate and centroid-free Z-coordinate of all points, divided by M.

[0051] Step S142: Perform eigenvalue decomposition on the point cloud spatial scatter matrix to solve for all eigenvalues ​​of the point cloud spatial scatter matrix and the eigenvector corresponding to each eigenvalue. Arrange the solved eigenvalues ​​in descending order of numerical value to generate a descending sequence of eigenvalues. Arrange each eigenvector in the order of its corresponding eigenvalue to form a eigenvector sequence. Extract the smallest eigenvalue from the descending sequence of eigenvalues. Extract the smallest eigenvector corresponding to the smallest eigenvalue from the eigenvector sequence. Determine the direction pointed to by the smallest eigenvector as the minimum spatial scattering direction of the independent connected component.

[0052] The Jacobi iterative method is used to perform eigenvalue decomposition on the 3x3 point cloud spatial scatter matrix. This yields three eigenvalues ​​and their corresponding three eigenvectors, each a 3x1 column vector. The three eigenvalues ​​are sorted in descending order of value to obtain the maximum, middle, and minimum eigenvalues. The eigenvector corresponding to the minimum eigenvalue represents the direction with the minimum scattering of the point cloud in 3D space; this direction is then determined as the minimum spatial scattering direction for this independent connected component.

[0053] Step S143: Take the minimum spatial distribution direction as the main extension direction of the independent connected component, whereby the main extension direction represents the narrowest dimensional direction of the point cloud spatial distribution within the independent connected component.

[0054] Normalize the eigenvector corresponding to the smallest eigenvalue to obtain the unit vector d, which is the principal extension direction of this independent connected component. The principal extension direction points to the narrowest dimension of the point cloud distribution.

[0055] Step S144: Construct a reference horizontal plane normal vector, the direction of which is perpendicular to the reference horizontal plane and points in the direction of increasing elevation, and calculate the angle between the main extension direction and the reference horizontal plane normal vector.

[0056] The reference horizontal plane normal vector is taken as n, with components (0, 0, 1). The angle θ between the main extension direction d and n is calculated using the inverse cosine function, θ = arccos(|d·n|), where · represents the vector dot product operation and || represents taking the absolute value. The value of θ ranges from 0 degrees to 90 degrees.

[0057] Step S145: Compare the directional angle with a preset acute angle threshold. When the directional angle is less than the preset acute angle threshold, mark the independent connected component as a wire component. When the directional angle is greater than or equal to the preset acute angle threshold, mark the independent connected component as a non-wire component.

[0058] Set a preset acute angle threshold θ0. If θ is less than θ0, it indicates that the angle between the main extension direction and the horizontal plane is small and close to horizontal, and the independent connected component is marked as a wire component. If θ is greater than or equal to θ0, it indicates that the angle between the main extension direction and the horizontal plane is large, and the independent connected component is marked as a non-wire component.

[0059] Step S146: Perform main extension direction consistency verification on the independent connected components marked as wire components, extract the projection coordinate components of each scanning point in the wire component along the main extension direction, and calculate the distribution span length of the projection coordinate components and the distribution center position of the projection coordinate components.

[0060] For an independent connected component labeled as a wire assembly, let its main extension direction be d, where d is a three-dimensional unit vector. The centroid coordinates of the u-th scan point within the component are (Xu, Yu, Zu). The projected coordinate components tu = Xu·dx + Yu·dy + Zu·dz, where dx, dy, and dz are the three components of d. Calculate the maximum value tmax and minimum value tmin of all tu. The distribution span length Lt = tmax - tmin. The distribution center position tc = (tmax + tmin) / 2.

[0061] Step S147: Based on the distribution span length of the projected coordinate components and the distribution center position, combined with the cross-sectional point cloud distribution range of the conductor assembly on the plane perpendicular to the main extension direction, construct the spatial extension direction envelope of the conductor assembly, perform morphological regularity analysis on the spatial extension direction envelope, calculate the cross-sectional area change curve of the spatial extension direction envelope along the main extension direction, extract the fluctuation amplitude parameter of the cross-sectional area change curve, and adjust the conductor type confidence of the conductor assembly based on the fluctuation amplitude parameter.

[0062] Divide Lt uniformly into Q segments, each segment having a length of Lt / Q. The t-value interval corresponding to the q-th segment is [tc-Lt / 2+(q-1)Lt / Q, tc-Lt / 2+qLt / Q]. Extract all scan points whose projected coordinates fall within this interval, and calculate the area Aq of the convex hull formed by these points on a plane perpendicular to d. Obtain the sequence [A1, A2, ..., AQ] along the d direction. The fluctuation amplitude parameter Rflu=(Amax-Amin) / Aavg, where Amax is the maximum value of the sequence, Amin is the minimum value of the sequence, and Aavg is the average value of the sequence. If Rflu is less than a preset threshold, the confidence level of the conductor type is increased; if Rflu is greater than the preset threshold, the confidence level of the conductor type is decreased.

[0063] Step S150: Perform cubic spline interpolation spatial curve fitting on the conductor assembly to fit the discrete point cloud into a smooth catenary parametric equation and output the three-dimensional spatial trajectory parameters of the conductor.

[0064] Step S151: For the conductor assembly, extract the three-dimensional spatial coordinate information of all conductor scanning points within the conductor assembly, project and sort all conductor scanning points along the main extension direction of the conductor assembly to obtain an ordered sequence of conductor scanning points arranged in spatial order, extract the first and last conductor scanning points from the ordered sequence of conductor scanning points, take the first conductor scanning point as the curve fitting start point and the last conductor scanning point as the curve fitting end point, and construct the total arc length parameter interval of the conductor spatial extension between the curve fitting start point and the curve fitting end point.

[0065] Extract the coordinates of all S scanning points within the conductor assembly. Calculate the projection value tv = Xvdx + Yvdy + Zvdz for the v-th scanning point. Sort the points by tv in ascending order to obtain an ordered sequence. The first point is the starting point of the curve fitting, and the last point is the ending point. The sum of the Euclidean distances between adjacent ordered points is the total arc length Stotal. The arc length parameter range is [0, Stotal].

[0066] Step S152: Divide the total arc length parameter interval evenly into multiple arc length parameter segment intervals, and select multiple continuous conductor scanning points in each arc length parameter segment interval as the fitting control point set of that arc length parameter segment interval.

[0067] Divide the area [0, Statotal] into U segments, each segment having a length of Statotal / U. Within the u-th segment, select all ordered scan points whose projected values ​​fall within that segment as the set of fitting control points.

[0068] Step S153: For each arc length parameter segment interval, construct a cubic spline interpolation basis function using the set of fitting control points within the arc length parameter segment interval. The cubic spline interpolation basis function satisfies the function value continuity condition, the first derivative continuity condition, and the second derivative continuity condition at the segment connection point of adjacent arc length parameter segment intervals.

[0069] Construct a cubic polynomial on the u-th segment interval [s(u), s(u+1)]: Xu(s) = au0 + au1(s-su) + au2(s-su)² + au3(s-su)³. Yu(s) and Zu(s) are similarly constructed, corresponding to the coefficients bu and cu respectively. Adjacent segments satisfy: Xu(s(u+1)) = X(u+1)(s(u+1)), Xu'(s(u+1)) = X(u+1)'(s(u+1)), Xu''(s(u+1)) = X(u+1)''(s(u+1)). Natural boundary conditions at both ends: X1''(0) = 0, XU''(Stotal) = 0. The Y and Z components are constructed similarly.

[0070] Step S154: Simultaneously establish the cubic spline interpolation basis functions for all segmented intervals of the arc length parameters to generate the global cubic spline interpolation curve equation system of the conductor assembly. Solve the global cubic spline interpolation curve equation system to obtain the complete cubic spline interpolation space curve of the conductor assembly. Convert the continuous trajectory expression of the complete cubic spline interpolation space curve in three-dimensional space into the form of the conductor catenary parametric equation. The conductor catenary parametric equation includes the catenary sag parameter, the catenary horizontal tension coefficient, and the spatial position parameter of the catenary suspension point.

[0071] By simultaneously solving the cubic polynomial coefficients of all U segments, a tridiagonal linear system of equations is obtained regarding the X-coordinate component coefficients au. The entire system of au is solved using the pursuit method. Similarly, bu and cu are solved to obtain the complete space curve (X(s), Y(s), Z(s)). This is then transformed into the catenary parametric equation form: Z(s) = a·cosh((sb) / a) + c, where a is the horizontal tension coefficient, b is the arc length of the center of symmetry, and c is the elevation offset of the lowest point of sag. a, b, and c are solved using the least squares fitting method. The spatial positions of the suspension points are (X(0), Y(0), Z(0)) and (X(Stotal), Y(Stotal), Z(Stotal)).

[0072] Step S155: Extract the three-dimensional spatial trajectory parameters of the conductor assembly from the catenary parameter equation of the conductor. The three-dimensional spatial trajectory parameters of the conductor include the continuous path coordinate sequence of the conductor in three-dimensional space, the spatial curvature value of the conductor at each arc length parameter position, and the spatial deflection value of the conductor at each arc length parameter position.

[0073] Sampling is performed along the arc length with a step size Δs to obtain a sampling point sequence (X(sn), Y(sn), Z(sn)), where sn = nΔs. At sn, the spatial curvature κ(sn) = |r'(sn)×r''(sn)| / |r'(sn)|³ is calculated, where r(s) = (X(s), Y(s), Z(s)), r' and r'' are the first and second derivative vectors respectively, and × denotes the cross product. The spatial torsion τ(sn) = (r'(sn)×r''(sn))·r'''(sn) / |r'(sn)×r''(sn)|², where · denotes the dot product.

[0074] Step S156: Perform catenary morphological feature decomposition on the three-dimensional spatial trajectory parameters of the conductor, and decompose the continuous path coordinate sequence in the three-dimensional spatial trajectory parameters of the conductor into a two-dimensional planar trajectory on the horizontal projection plane and a sag variation curve in the vertical projection direction.

[0075] Projecting (X(sn), Y(sn), Z(sn)) onto the XY plane yields a two-dimensional trajectory (X(sn), Y(sn)). The Z(s) curve is extracted as the overhang variation curve.

[0076] Step S157: Extract the horizontal azimuth sequence and horizontal deflection radius sequence of the conductor assembly based on the two-dimensional plane trajectory, and extract the position of the maximum sag amplitude and the position of the sag symmetry center of the conductor assembly based on the sag variation curve.

[0077] Calculate the angle between the tangent direction and true north for each point on the two-dimensional trajectory to obtain the azimuth sequence α(sn). Calculate the reciprocal of the radius of curvature at each point to obtain the deflection radius sequence R(sn). Differentiate Z(s) and find the arc length smax where Z'(s) = 0 and Z''(s) > 0, which is the position of maximum overhang amplitude. The center of symmetry is sb = b.

[0078] Step S158: Combine the horizontal azimuth sequence, the horizontal deflection radius sequence, the maximum overhang amplitude position, the overhang symmetry center position, and the spatial curvature and spatial deflection values ​​in the three-dimensional spatial trajectory parameters of the conductor to generate a complete set of conductor spatial morphology description parameters, and output the set of conductor spatial morphology description parameters.

[0079] The combined output data structure consists of records [sn, X(sn), Y(sn), Z(sn), α(sn), R(sn), κ(sn), τ(sn)] corresponding to each arc-length sampling point sn, along with global parameters smax, sb, a, b, c. This data is then output for further processing or storage.

[0080] Step S210: The method further includes: performing tower head structure identification processing on the tower assembly, extracting a tower head point cloud subset and a tower body point cloud subset from the tower assembly, wherein the tower head point cloud subset corresponds to the spatial point cloud of the top crossarm and hanging point of the tower, and the tower body point cloud subset corresponds to the spatial point cloud of the main material and diagonal material of the tower.

[0081] A layered analysis of the tower components along the elevation direction is performed. The horizontal projection distribution range of the point cloud within each elevation layer is calculated. The point cloud in the tower body is continuous along the elevation direction, and the horizontal projection range decreases with increasing elevation. The point cloud at the tower head suddenly increases in horizontal projection range within a certain elevation range, showing a laterally extending cluster of point clouds. The boundary elevation Hdiv of the abrupt change in projection range along the elevation direction is identified; points above Hdiv are the subset of the tower head point cloud, and points below Hdiv are the subset of the tower body point cloud.

[0082] Step S211: Perform spatial location processing on the tower head point cloud subset to locate the hanging point, search for local point cloud clusters extending horizontally in the tower head point cloud subset, and extract the spatial geometric center coordinates of each local point cloud cluster as the candidate hanging point spatial location.

[0083] Within the point cloud subset at the tower head, a density clustering algorithm based on Euclidean distance is used to search for local point cloud clusters. The cluster radius is set to 1.5 times the typical size of the conductor crossarm component. For each cluster, the average values ​​of the X, Y, and Z coordinates of all scanned points within it are calculated to form the spatial geometric center coordinates of the cluster, which serve as the spatial location of candidate wire hanging points.

[0084] Step S212: Calculate the horizontal spatial distance between adjacent candidate wire hanging points, verify the rationality of the horizontal spatial distance according to the standard phase-to-phase distance range in the transmission line design specification, eliminate candidate wire hanging point spatial positions that do not conform to the standard phase-to-phase distance range, and obtain the final wire hanging point spatial position.

[0085] Candidate connection points are sorted by their X-axis coordinates. The horizontal distance between adjacent candidate points is calculated as the absolute value of the difference between the X-coordinates of two adjacent points. This horizontal distance is compared with the standard phase-to-phase distance range, which has a lower and upper limit determined by the voltage level. If the horizontal distance falls between the lower and upper limits, the pair of candidate points is retained; otherwise, the candidate point with the higher elevation is discarded. The retained candidate points constitute the final spatial location of the connection points.

[0086] Step S213: Based on the spatial location of the final hanging point and the spatial location of the conductor endpoint of the conductor assembly, perform spatial distance association matching, establish a connection correspondence between the conductor endpoint with the shortest spatial distance and the hanging point, and determine the hanging position of each conductor assembly on the tower.

[0087] Extract the three-dimensional coordinates of the beginning and end of each conductor assembly as the spatial location of the conductor endpoint. For each final attachment point, calculate its three-dimensional Euclidean distance to all conductor endpoints. Establish a pairing relationship between the conductor endpoint with the smallest distance and the attachment point. This attachment point is the attachment position of the conductor assembly on the tower.

[0088] Step S214: Analyze the conductor arrangement sequence of the conductor assemblies associated with each splicing position on the same tower, extract the arrangement direction and spacing of each conductor assembly on the horizontal projection plane, and generate conductor phase sequence arrangement layout information.

[0089] Project all connection points on the same tower onto the XY horizontal plane. Perform principal component analysis on the projected points; the direction of the first principal component is the conductor arrangement direction. Calculate the spacing between adjacent projected points along the arrangement direction to obtain the arrangement spacing sequence. Sort and number the conductor components according to their arrangement direction coordinates from smallest to largest to generate conductor phase sequence arrangement layout information.

[0090] Step S310: The method further includes: performing conductor morphology change analysis on the conductor assembly between two adjacent towers within the span, extracting conductor spatial position sampling points at equal arc length intervals along the conductor spatial extension direction, and calculating the spatial distance increment and direction change angle between adjacent conductor spatial position sampling points.

[0091] Extract the catenary parametric equations for the conductor assemblies within the span between two adjacent towers. Obtain W sampling points at equal intervals along the arc length parameter; each sampling point contains three-dimensional coordinates and the arc length parameter. Calculate the three-dimensional Euclidean distance between adjacent sampling points to obtain a distance increment sequence. Calculate the angle between the tangent directions at adjacent sampling points; the tangent direction is the direction of the first derivative vector of the catenary parametric equation, and the angle is obtained by taking the inverse cosine of the dot product of the two tangent direction vectors.

[0092] Step S311: Input the spatial distance increment and the directional change angle into the catenary tension state inversion model, and calculate the real-time horizontal tension distribution of the conductor assembly within the span to obtain the tension change curve along the conductor extension direction.

[0093] The catenary tension state inversion model is constructed based on the static equilibrium equation of the conductor. The horizontal tension at any point on the conductor is equal to the weight per unit length of the conductor multiplied by the radius of curvature at that point, divided by the sine of the angle between the tangent direction at that point and the horizontal plane. The radius of curvature is obtained by taking the reciprocal of the spatial curvature at that point, which has already been calculated in step S155. By substituting the radius of curvature and the tangent direction angle at each sampling point, the horizontal tension value at each point is calculated, generating a tension variation curve along the arc length direction.

[0094] Step S312: Extract the tension peak position and tension valley position based on the tension change curve, and perform spatial correspondence analysis between the tension peak position and the spatial coordinates of the terrain protrusion point in the micro-topography point cloud data to identify the local influence area of ​​the terrain protrusion on the conductor tension.

[0095] Traverse the tension variation curve to search for local maxima and minima. At a local maximum, the tension value is greater than the tension values ​​of its immediate neighbors; at a local minimum, the tension value is less than the tension values ​​of its immediate neighbors. Extract the 3D coordinates of the tension peak location and calculate the spatial distance between them and the 3D coordinates of the terrain protrusions in the micro-topography point cloud. If the horizontal projection distance between a tension peak location and a terrain protrusion is less than a preset correlation distance, the terrain protrusion is determined to have a local influence on the tension peak. Areas with this correlation are marked as locally affected areas.

[0096] Step S313: Perform safety spacing verification on the conductor assemblies within the local influence area, calculate the vertical distance between the conductor spatial location sampling point and the point cloud on the terrain surface below, compare the vertical distance with the minimum safety distance specified in the transmission line safety regulations, and mark the conductor safety risk sections with insufficient vertical distance.

[0097] Within the local influence area, for each traverse sampling point, the topographic surface point cloud within a certain horizontal range directly below it is searched. The topographic surface point cloud is extracted from the ground point cloud of P0. The vertical distance between the traverse sampling point and the topographic point below it is calculated, taking the absolute value of the difference between the Z-coordinate of the traverse sampling point and the Z-coordinate of the topographic point. This vertical distance is compared with the minimum safe distance, which is obtained from a table in the regulations based on the voltage level. If the vertical distance is less than the minimum safe distance, the arc length segment containing the traverse sampling point is marked as a safety risk segment.

[0098] Step S314: Integrate the spatial location information of the safety risk section of the conductor, the tension value of the corresponding section in the tension change curve, and the spatial coordinates of the topographic protrusion point to generate a conductor tension-topographic associated risk distribution map.

[0099] For each safety risk section, record its starting and ending arc length parameters, starting and ending three-dimensional coordinates, average tension value, maximum tension value within the section, and coordinates of the associated topographic convexity point. Integrate the above information according to spatial location to generate a conductor tension topographic associated risk distribution map.

[0100] Step S410: The method further includes: performing wind-sensitive section identification processing on the conductor assembly, extracting the local bending curvature change drastic sections of the conductor assembly in the catenary morphology, and calculating the curvature change gradient amplitude of each local bending curvature change drastic section.

[0101] The spatial curvature values ​​at each sampling point are extracted from the catenary parametric equation of the conductor assembly, as calculated in step S155. The curvature values ​​are then subjected to first-order difference along the arc length direction to obtain a sequence of curvature change rates. Continuous arc length intervals where the absolute value of the curvature change rate exceeds a preset threshold are searched, and these intervals are marked as segments with drastic local curvature changes. Within each drastic segment, the curvature change gradient magnitude is calculated; the gradient magnitude is equal to the maximum absolute value of the curvature change rate within that segment.

[0102] Step S411: Compare the curvature change gradient magnitude with a preset curvature change gradient threshold, and mark the segment where the curvature change gradient magnitude is greater than the preset curvature change gradient threshold as a morphological change sensitive segment. The morphological change sensitive segment corresponds to the position in the catenary morphology where the degree of curvature changes drastically.

[0103] The preset curvature change gradient threshold is determined by the conductor type and span. The gradient amplitude of each abrupt change section is compared with this threshold. Sections with amplitudes greater than the threshold are marked as morphological change sensitive sections. Morphological change sensitive sections are located at locations in the catenary morphology where the degree of curvature changes drastically, corresponding to the vicinity of clamp exits, vibration damper installation points, or conductor reversal points.

[0104] Step S412: Perform conductor cross-section orientation analysis on the morphological change sensitive section, extract the cross-section normal vector of the conductor at each sampling position in the morphological change sensitive section, and calculate the incident angle between the cross-section normal vector and the prevailing wind direction.

[0105] At each sampling point within the morphologically sensitive section, the conductor cross-section normal vector is taken as the tangent direction of the catenary parametric equation at that point, as calculated in step S311. The prevailing wind direction is obtained from meteorological statistics of the transmission corridor area and represented as a three-dimensional unit vector. The incident angle is equal to the absolute value of the angle between the cross-section normal vector and the prevailing wind direction vector, calculated by taking the inverse cosine of the absolute value of the dot product of the two vectors.

[0106] Step S413: If the incident angle exceeds the preset angle threshold and the magnitude of the curvature change gradient exceeds the preset curvature change gradient threshold, then the sensitive section of the morphological change is identified as the target risk section.

[0107] The preset angle threshold is set as the upper limit of the angle range where conductor vibration is prone to occur. For sections sensitive to abrupt changes in morphology, both the incident angle and the magnitude of the curvature change gradient are considered. If both conditions are met, the section is designated as the target risk section. The target risk section is a high-risk location where conductor fatigue and strand breakage can occur due to wind vibration.

[0108] Step S414: Perform local point cloud density enhancement sampling on the point cloud of the conductor within the target risk section, and repeatedly scan the target risk section from multiple angles using the UAV-borne lidar to obtain supplementary point cloud data for the target risk section.

[0109] The spatial coordinates of the target risk area are output to the UAV flight control system. The UAV performs a multi-angle repeated scan of this area, with the scanning angle varying in preset steps within a range of positive and negative preset angles away from the nadir. The point cloud data acquired in each scan is registered and fused with the original point cloud to increase the point cloud density of this area to a preset multiple or more than the original density.

[0110] For example, in step S510: the method further includes: performing multi-temporal point cloud change detection processing on the overhead point cloud set, loading multi-period transmission corridor point cloud data sets with different acquisition time phases, and unifying the transmission corridor point cloud data sets with different time phases to the same spatial reference coordinate system.

[0111] Two sets of point cloud data from the power transmission corridor, acquired at different times, are loaded and denoted as P10 and P20, respectively. The acquisition time for P10 is T1, and for P20 it is T2, with T2 being later than T1. The spatial reference coordinate systems of the two point clouds may differ. An iterative nearest-point algorithm is used to register P20 to the coordinate system of P10. During registration, the tower point cloud is used as the control reference, ensuring that the tower point cloud does not undergo spatial position changes in the short term. After registration, the two point clouds are in the same coordinate system.

[0112] Step S511: Perform the elevation histogram adaptive ground filtering and the voxelized spatial connectivity clustering processing on the point cloud data sets of different time phases of the transmission corridor after unified spatial reference, respectively, to generate the corresponding conductor components for each time.

[0113] Perform all the processes of steps S120 to S150 on P10 to generate the conductor assembly set and conductor three-dimensional spatial trajectory parameters for phase T1. Perform all the above processes on P20 in the same way to generate the conductor assembly set and conductor three-dimensional spatial trajectory parameters for phase T2.

[0114] Step S512: Spatial registration of conductor components at different time phases, differential comparison of the three-dimensional spatial trajectory parameters of the same conductor at different time phases, and calculation of the displacement change of the conductor in the vertical direction and the offset change in the horizontal direction.

[0115] By matching the spatial locations of the pole and tower stringing points, the correspondence between the two phases of traverse components is determined, establishing the relationship between the same traverse in the two phases of data. For the same traverse, the three-dimensional coordinates of both phases are extracted along the arc length parameter direction using the same sampling point locations. The vertical displacement change is equal to the difference between the Z-coordinate of phase T2 and the Z-coordinate of phase T1. The horizontal offset change is equal to the Euclidean distance between the XY-plane projected coordinates of the two phases.

[0116] Step S513: Construct a spatiotemporal deformation sequence of the conductor based on the displacement change and the offset change, and extract the trend features of the conductor deformation over time from the spatiotemporal deformation sequence. The trend features include the deformation direction consistency parameter and the deformation rate change pattern.

[0117] For each sampling point, the displacement change and offset change are calculated, and a deformation sequence is constructed along the arc length direction. The deformation direction consistency parameter is taken as the average cosine similarity between the horizontal offset direction vectors of each sampling point. The deformation rate change pattern is taken as the trend of the rate value obtained by dividing the displacement change by the time interval (T2-T1) along the arc length direction, and the slope of the rate change is obtained by linear fitting.

[0118] Step S514: Perform correlation analysis between the trend features and the terrain change areas in the micro-topography point cloud data to determine the location of the terrain deformation source area that causes the conductor deformation and the direction of the mechanical action of the terrain deformation source area on the conductor.

[0119] Topographic change regions are extracted from two phases of ground point clouds. These regions are defined as those where the absolute value of the difference in ground elevation between the two phases exceeds a preset threshold. The deformation direction of the traverse line is compared with the spatial location of the topographic change regions. If the traverse line's deformation direction points to a topographic change region, and the direction of ground elevation change in that region is consistent with the traverse line's displacement direction, then that topographic change region is identified as a topographic deformation source region. The direction of the mechanical action of the topographic deformation source region on the traverse line is determined by the direction of the ground elevation change gradient in that region.

[0120] Step S610: The method further includes: performing pole tilt state assessment processing on the pole assembly, extracting the point cloud of the vertical main material direction of the pole from the pole assembly, and performing spatial straight line fitting on the point cloud of the vertical main material direction of the pole to obtain the direction vector of the pole central axis.

[0121] Vertical main material direction point clouds are extracted from the tower body point cloud subset of the tower assembly. A random sampling consensus algorithm is used for spatial line fitting. Two points are randomly selected to determine a candidate line. The distance from all points in the tower body point cloud to this candidate line is calculated, and points with a distance less than a preset distance threshold are recorded as interior points. After multiple iterations, the candidate line with the most interior points is selected as the fitting result. The direction vector of the fitted line is the direction vector of the tower's central axis.

[0122] Step S611: Calculate the spatial tilt angle between the direction vector of the tower's central axis and the vertical reference vector, and compare the spatial tilt angle with the allowable tilt angle threshold in the transmission line tower operation regulations to obtain the tower tilt state assessment result.

[0123] The vertical reference vector is taken as (0, 0, 1). The spatial tilt angle is equal to the angle between the tower's central axis direction vector and the vertical reference vector, calculated by taking the inverse cosine of the absolute value of the dot product of the two vectors. This spatial tilt angle is compared with the allowable tilt angle threshold in the operating procedures, the threshold being determined by the tower type and voltage level. If the spatial tilt angle is less than the threshold, it is assessed as normal; otherwise, it is assessed as excessive tilt.

[0124] Step S612: Perform insulator string point cloud segmentation processing on the tower assembly, separate the insulator string connection point cloud from the tower head point cloud subset of the tower assembly, and extract the spatial suspension direction vector of the insulator string.

[0125] Within the point cloud subset at the tower head, there exists a slender point cloud structure distributed vertically within a certain distance below the suspension point; this is the insulator string point cloud. Extracting this slender point cloud structure and performing principal component analysis on its internal point cloud reveals that the direction of the first principal component is the spatial suspension direction vector of the insulator string.

[0126] Step S613: Calculate the relative angle between the spatial suspension direction vector of the insulator string and the direction vector of the central axis of the tower, determine the wind deflection state of the insulator string relative to the tower based on the change of the relative angle, and generate wind deflection assessment information of the insulator string.

[0127] The relative angle is equal to the angle between the suspension direction vector of the insulator string and the direction vector of the tower's central axis. Ideally, the insulator string should be vertically suspended, with the suspension direction vector aligned with the vertical direction. Under wind-induced deviation, the insulator string deviates from the vertical direction. The difference between the relative angle and the ideal vertical angle is calculated; a larger difference indicates a more severe wind-induced deviation, generating wind-induced deviation assessment information for the insulator string.

[0128] Step S614: The pole tilt state assessment result is fused with the insulator string wind deflection assessment information to generate a comprehensive attitude assessment report of the pole and hanging hardware.

[0129] The tower tilt angle and insulator string offset angle are weighted and fused according to a preset weighting coefficient to obtain a comprehensive attitude anomaly index. Based on the numerical range of the comprehensive attitude anomaly index, an assessment level (normal, warning, or abnormal) is output, generating a comprehensive attitude assessment report for the tower and its wiring fittings.

[0130] Step S710: The method further includes: performing three-dimensional reconstruction processing on the conductor assembly to determine the conductor icing morphology; extracting a subset of the conductor surface point cloud from the conductor assembly; performing cylindrical surface fitting processing on the subset of the conductor surface point cloud to obtain a reference cylindrical surface geometric model of the conductor in the non-icing state; calculating the spatial distance between all conductor scanning points in the conductor assembly and the reference cylindrical surface geometric model; and extracting scanning points located outside the reference cylindrical surface geometric model and whose spatial distance is greater than a preset icing determination distance to form a conductor icing surface point cloud set.

[0131] The conductor assembly is segmented along its arc length, and the point cloud of the conductor surface within each segment is fitted using a cylindrical surface. The axis of the cylindrical surface is taken as the tangent direction of the catenary parametric equation at the center of the segment, and the radius of the cylinder is determined by least squares fitting. The fitted cylindrical surface is the reference cylindrical surface geometric model for the conductor in the icy-free state. The radial distance from all scan points within the conductor assembly to the reference cylindrical surface is calculated. If the radial distance of a scan point is greater than the preset icing determination distance, then that point is included in the conductor icy surface point cloud set.

[0132] Step S711: Perform triangular mesh surface reconstruction processing on the point cloud set of the ice-covered surface of the conductor to generate a triangular mesh model of the ice-covered surface of the conductor. The triangular mesh model of the ice-covered surface of the conductor contains a continuous surface topology of the outer contour of the ice.

[0133] A Poisson surface reconstruction algorithm is used to reconstruct triangular mesh surfaces from the point cloud set of the ice-covered surface of a conductor. An octree spatial index structure is constructed, the normal vector field of the point cloud is calculated, the Poisson equation is solved to obtain the indicator function, and the isosurface of the indicator function is extracted as the triangular mesh surface. The generated triangular mesh model includes vertex coordinates and the topological connectivity of triangular patches.

[0134] Step S712: Perform cross-sectional slicing on the triangular mesh model of the icing surface of the conductor along the catenary spatial extension direction to obtain multiple icing cross-sectional contour lines perpendicular to the conductor extension direction.

[0135] Cross-sectional positions are taken at equal intervals along the arc length parameter direction. At each cross-sectional position, the intersection of a plane perpendicular to the catenary tangent direction with the triangular mesh model is obtained to obtain the ice-covered cross-sectional profile at that position. The profile is represented as an ordered sequence of two-dimensional vertex coordinates.

[0136] Step S713: Compare the morphological differences of adjacent ice-covered cross-sectional contours, extract the difference magnitude of adjacent ice-covered cross-sectional contours in three morphological dimensions: the major axis direction, the minor axis direction, and the contour perimeter, and mark the transition area of ​​adjacent cross-sections where the difference magnitude in all three morphological dimensions exceeds the preset morphological difference threshold as the ice-covered morphological abrupt transition area.

[0137] For each icing cross-sectional profile, calculate the angle along the major axis, length of the major axis, length of the minor axis, and perimeter of the equivalent ellipse. For adjacent cross-sectional profiles, calculate the angle difference along the major axis, the relative rate of change of the major axis length, the relative rate of change of the minor axis length, and the relative rate of change of the perimeter. If the rate of change in all three dimensions exceeds their respective preset thresholds, the transition section between the adjacent cross-sections is marked as an abrupt transition zone in icing morphology.

[0138] Step S714: Extract the eccentricity direction vector of the geometric center of each ice-covered cross-section contour line relative to the geometric center of the conductor cross-section, and perform directional continuity tracking of the eccentricity direction vectors of adjacent ice-covered cross-section contour lines to obtain the rotational winding trajectory of the eccentricity direction along the extension direction of the conductor.

[0139] The geometric center of the conductor cross-section is taken as the point on the axis of the reference cylinder corresponding to the arc length. The geometric center of the ice-covered cross-section profile is taken as the arithmetic mean of the coordinates of the vertices of the profile. The eccentric direction vector points from the geometric center of the conductor cross-section to the geometric center of the ice-covered cross-section and is projected onto the cross-sectional plane. By tracking the angular change of the eccentric direction vector section by section along the arc length, the rotational winding trajectory of the eccentric direction along the extension direction of the conductor is obtained.

[0140] Step S715: Perform an angle correlation analysis between the major axis direction of the ice-covered cross-sectional profile and the eccentric direction vector to determine the dominant azimuth sector of ice accumulation in the circumferential direction of the conductor, and record the continuous change trend of the dominant azimuth sector along the extension direction of the conductor as the migration path of ice accumulation azimuth.

[0141] Calculate the angle between the major axis and the eccentric direction for each cross section. If the angle is less than a preset small angle threshold, it indicates that icing mainly accumulates along the eccentric direction. Statistically analyze the dominant azimuth angles of icing accumulation on each cross section, and determine the azimuth interval with the highest frequency in the angle distribution as the dominant eccentric sector. Track the continuous change of the center angle of this sector along the arc length direction and record it as the migration path of the icing azimuth.

[0142] Step S716: Spatially superimpose the location information of the abrupt transition zone of the icing morphology, the rotation and winding trajectory of the eccentric direction along the extension direction of the conductor, and the migration path of the icing aggregation azimuth to generate a three-dimensional icing morphology distribution model of the conductor.

[0143] The starting and ending arc length parameters of the icing morphology transition zone, the angle sequence of the eccentric rotation and winding trajectory, and the angle sequence of the icing aggregation azimuth migration path are uniformly aligned according to the arc length parameters and superimposed onto the three-dimensional spatial trajectory of the conductor catenary to generate a three-dimensional icing morphology distribution model of the conductor.

[0144] Step S717: Spatially superimpose the three-dimensional morphological distribution model of the ice-covered conductor with the spatial trajectory parameter equation of the catenary of the conductor, extract the symmetry characteristics of the distribution of the ice-covered morphological transition zone on both sides of the lowest point of the catenary sag, and determine whether the ice-covered conductor meets the galloping excitation conditions based on the symmetry characteristics of the distribution.

[0145] Extract the arc length smax of the lowest point of the catenary sag. Count the number of abrupt transition zones in icing morphology within each preset arc length range to the left and right of smax. The symmetry characteristic of the distribution is taken as the ratio of the number on the left to the number on the right. If this ratio deviates from 1 by more than a preset symmetry threshold, it indicates that the icing distribution is asymmetrical and meets the conditions for galloping.

[0146] Step S718: When the distribution symmetry of the transition zone of icing morphology change on both sides of the lowest point of the catenary sag of the conductor is lower than the preset symmetry threshold, the corresponding conductor segment is marked as the icing dance sensitive segment, and the spatial location information of the icing dance sensitive segment and the migration path of the icing clustering direction are output.

[0147] The arc length segments of the conductor with a distribution symmetry below the threshold are marked as ice-gale-sensitive segments. The starting and ending arc length parameters, starting and ending three-dimensional coordinates, ice aggregation azimuth migration path, and location information of the ice morphology transition zone of this segment are output.

[0148] Step S810: The method further includes: performing a conductor micro-vibration fatigue sensitive point identification process on the conductor assembly, extracting the spatial curvature numerical sequence of the conductor at each arc length parameter position from the conductor catenary spatial trajectory parameter equation of the conductor assembly, performing point-by-point curvature change rate analysis on the spatial curvature numerical sequence along the extension direction of the conductor, searching for local arc length intervals where the curvature change rate shows alternating positive and negative oscillations, and marking the local arc length intervals where the alternating frequency of the curvature change rate exceeds a preset alternating frequency threshold as curvature oscillation dense areas.

[0149] From the spatial curvature numerical sequence output in step S155, calculate the rate of curvature change between adjacent sampling points. Search for continuous arc-length intervals where the signs of the rate of curvature change alternate between positive and negative. Count the number of alternations per unit arc length within each alternation interval as the alternation frequency. Mark arc-length intervals with alternation frequencies greater than a preset threshold as areas of dense curvature oscillation.

[0150] Step S811: Extract the conductor point cloud subset corresponding to the curvature oscillation dense region from the conductor assembly, perform conductor surface normal vector estimation processing on the conductor point cloud subset, and obtain the surface normal vector direction of each scanning point on the conductor surface within the curvature oscillation dense region.

[0151] The scanning points of the conductor within the arc length range of the dense curvature oscillation region are extracted to form a subset of the point cloud. For each scanning point in the point cloud subset, its spatial neighborhood is searched, and local plane fitting is performed using the neighborhood points. The normal vector of the fitted plane is the surface normal vector direction of the scanning point.

[0152] Step S812: Perform orthogonality analysis on the surface normal vector direction of each scanning point on the conductor surface within the curvature oscillation dense region and the tangent direction of the catenary trajectory at that scanning point. Select scanning points whose angle between the surface normal vector direction and the tangent direction deviates from the orthogonal relationship to form a set of geometric anomalies on the conductor surface. Perform spatial clustering processing on the set of geometric anomalies on the conductor surface. Aggregate surface geometric anomalies with a spatial distance smaller than the preset clustering neighborhood radius into anomaly clusters and extract the spatial bounding box range of each anomaly cluster.

[0153] Calculate the angle between the surface normal vector of each scan point and the tangent direction of the catenary. The surface normal vector of a normal conductor should be orthogonal to the tangent direction, with an angle close to 90 degrees. If the angle deviates from 90 degrees by more than a preset deviation threshold, the scan point is marked as a geometric outlier. All geometric outliers are spatially clustered using a density-based clustering algorithm, with the cluster radius set to a preset neighborhood radius. Each cluster of points is then treated as an outlier cluster, and its minimum bounding box is calculated.

[0154] Step S813: Register the spatial bounding box range of each anomaly cluster with the spatial position of the conductor in the conductor catenary spatial trajectory parameter equation to determine the arc length interval position of each anomaly cluster on the conductor arc length parameter axis.

[0155] Projecting the center point of the spatial bounding box of an anomaly cluster onto the catenary of the conductor, the arc length parameter corresponding to the projection point is the position of the cluster on the arc length parameter axis. The arc length range of the cluster is taken as the arc length span of the bounding box projected along the tangent direction of the conductor.

[0156] Step S814: Perform principal component analysis on the geometric anomalies on the conductor surface within each anomaly cluster, extract the maximum and minimum scattering directions of the point cloud within the anomaly cluster on the tangential plane of the conductor surface, and determine whether the geometric morphology of the anomaly cluster exhibits a crack-like linear distribution based on the ratio of the scattering span of the maximum and minimum scattering directions. Mark the anomaly clusters exhibiting a crack-like linear distribution as candidate regions for microcracks on the conductor surface. Perform abrupt point cloud density detection on the neighboring regions of the candidate regions along the conductor extension direction. If there is a step change in point cloud density at the boundary of the candidate regions, then the candidate regions are identified as sensitive points for conductor fatigue cracks.

[0157] Within the cluster of anomalous points, the point cloud is projected onto the tangent plane of the conductor surface. Principal component analysis is performed on the projected points to obtain the first principal component direction (maximum dispersion direction) and the second principal component direction (minimum dispersion direction). The ratio of the maximum dispersion span to the minimum dispersion span is calculated. If this ratio is greater than a preset linear distribution threshold, it indicates that the point cloud within the cluster exhibits a crack-like linear distribution, and it is marked as a candidate region for microcracks on the conductor surface. The rate of change of point cloud density is calculated within the immediate vicinity of this candidate region along the conductor direction. If the rate of change of density at the boundary exceeds a preset step threshold, it indicates a sudden change in point cloud density, and this candidate region is confirmed as a sensitive point for conductor fatigue cracks.

[0158] Step S815: Extract the arc length interval where the conductor fatigue crack sensitive point is located, the azimuth angle of the conductor fatigue crack sensitive point on the circumference of the conductor cross section, and the curvature change amplitude of the curvature oscillation dense area corresponding to the conductor fatigue crack sensitive point, and combine them to generate a feature description of the conductor fatigue crack sensitive point.

[0159] The arc length interval is determined by the arc length parameter projected onto the catenary from the center of the fatigue crack sensitive point. The azimuth angle is determined by the angular coordinates of the sensitive point center on the conductor cross-section. The curvature change amplitude is determined by the absolute value of the maximum rate of curvature change within the corresponding dense curvature oscillation region. These parameters are combined to generate a characteristic description of the conductor fatigue crack sensitive point.

[0160] Step S816: Perform spatial proximity matching between the characteristic description of the conductor fatigue crack sensitive point and the known spatial coordinates of the suspension wire clamp outlet, the vibration damper installation position and the spacer clamping position to determine the spatial correlation between the conductor fatigue crack sensitive point and the hardware constraint boundary.

[0161] Extract the three-dimensional spatial coordinates of the suspension clamp outlet, vibration damper installation position, and spacer clamping position from the transmission line design data or point cloud. Calculate the three-dimensional spatial distance between each fatigue crack sensitive point and the aforementioned hardware constraint boundary position. If this distance is less than a preset proximity threshold, it is determined that there is a spatial correlation between the sensitive point and the hardware constraint boundary. Output a list of correlation relationships.

[0162] Step S910: The method further includes: performing a coupled wind vibration response analysis of the tower assembly and the conductor assembly, extracting the spatial position of the conductor suspension point from the tower head point cloud subset of the tower assembly, and connecting and matching the spatial position of each conductor suspension point with the spatial position of the conductor endpoint of the corresponding conductor assembly.

[0163] Extract the 3D coordinates of each wire hanging point from the cloud subset of tower head points of the tower assembly. Extract the 3D coordinates of the beginning and end points of each conductor assembly. Establish a one-to-one correspondence between the wire hanging points and the conductor endpoints through spatial distance minimization matching.

[0164] Step S911: Extract the structural skeleton line of the tower assembly. Extract the spatial positions of the main members corner points and main member connection nodes from the tower point cloud. Construct a tower spatial skeleton line model with the main members as skeleton units and the main member connection nodes as connection hinge points. Assign bending stiffness and torsional stiffness characteristics to each skeleton unit in the tower spatial skeleton line model according to the material section properties. Assign rigid connection constraints or semi-rigid connection constraints to each connection hinge point according to the node connection method. Generate a simplified tower mechanical model.

[0165] Corner points of the main tower members are extracted from the point cloud of the tower body, and an edge detection algorithm based on normal vector transformation is used. Adjacent corner points are connected to form main tower skeleton elements. For each skeleton element, the cross-sectional area, moment of inertia, and polar moment of inertia are assigned according to the tower design data, and the bending stiffness and torsional stiffness are calculated accordingly. For connecting hinge points, constraint types are assigned according to the connection method: rigid connections are assigned to main tower through nodes, and semi-rigid connections are assigned to auxiliary tower connection nodes. A simplified mechanical model of the tower, including node coordinates, element topology, cross-sectional properties, and constraint conditions, is generated.

[0166] Step S912: Discretize the parametric equation of the conductor assembly along the catenary spatial trajectory, and discretize the continuous conductor catenary into multiple conductor micro-segment units. The conductor micro-segment units are connected in series through tension transmission nodes.

[0167] The catenary conductor is discretized along its arc length into micro-segments of equal arc length. The positions of the two endpoints of each micro-segment are determined by the catenary parametric equations. Adjacent micro-segments share intermediate nodes, through which tension is transferred. Each micro-segment imparts mass per unit length and axial stiffness to the conductor.

[0168] Step S913: Connect the conductor endpoints of each conductor micro-segment unit to the conductor suspension point spatial position of the simplified tower mechanical model through the insulator string suspension point constraint, and establish the force transmission coupling boundary conditions between the conductor micro-segment unit and the tower skeleton unit.

[0169] The end node of the outermost micro-segment of the conductor is connected to the suspension point of the tower via an insulator string. The insulator string is simplified as a rigid rod element, with one end hinged to the end node of the conductor and the other end hinged to the suspension point of the tower. The hinged constraint transmits force but not bending moment.

[0170] Step S914: Apply a wind load spatial distribution field to the simplified mechanical model of the tower and the conductor micro-segment unit coupled thereto. The wind load spatial distribution field includes a wind speed profile that varies along the tower height and a fluctuating wind pressure distribution that varies along the conductor span.

[0171] The wind speed profile employs an exponential law model, where the wind speed at different heights is equal to the wind speed at the reference height multiplied by the power of the ratio of height to reference height. The power exponent is determined by the terrain roughness category. The fluctuating wind pressure distribution uses a simple harmonic wave superposition model along the span direction, with the frequency and amplitude of each order of simple harmonic waves determined based on meteorological statistical parameters. The wind load acting on the tower frame unit and conductor micro-segment unit is equal to the wind pressure multiplied by the windward area and then by the aerodynamic coefficient.

[0172] Step S915: Perform coupled dynamic response analysis on the simplified mechanical model of the tower and the conductor micro-segment unit after applying the wind load spatial distribution field, and output the displacement response time history of the connecting hinge point of the tower skeleton unit and the tension fluctuation time history of the tension transmission node of the conductor micro-segment unit.

[0173] The equations of motion for the coupled system are solved using Newmark's successive integration method. The time step and total computation time are set. Within each time step, the displacement, velocity, and acceleration of each node are calculated, and the internal forces of the elements are updated. The displacement time histories of each tower hinge point and the tension time histories of each conductor node are recorded. The calculation results are output as time series data.

[0174] Step S916: Extract the connection hinge points whose displacement amplitude exceeds a preset displacement threshold from the displacement response time history of the connection hinge points of the tower skeleton unit as tower wind vibration sensitive nodes. Extract the tension transmission nodes whose tension fluctuation amplitude exceeds a preset tension fluctuation threshold from the tension fluctuation time history of the tension transmission nodes of the conductor micro-segment unit as conductor wind vibration sensitive nodes. Perform spatial correspondence analysis between the spatial position of the tower wind vibration sensitive nodes in the tower spatial skeleton line model and the arc length position of the conductor wind vibration sensitive nodes in the conductor catenary spatial trajectory parameter equation to identify the spatial coupling region between tower wind vibration and conductor wind vibration.

[0175] For each tower hinge point, the displacement amplitude is calculated based on its displacement time history, where the amplitude is the maximum distance the displacement deviates from the initial position during the time history. The displacement amplitude is compared to a preset displacement threshold; nodes exceeding this threshold are marked as wind-sensitive nodes for tower vibration. For each conductor node, the tension fluctuation amplitude is calculated based on its tension time history, where the fluctuation amplitude is half the difference between the maximum and minimum tension values ​​during the time history. The fluctuation amplitude is compared to a preset tension fluctuation threshold; nodes exceeding this threshold are marked as wind-sensitive nodes for conductor vibration. The tower sensitive nodes and conductor sensitive nodes are spatially correlated through the connection relationships of the hanging points, identifying coupling regions with similar spatial locations.

[0176] Step S917: Perform a direction consistency analysis on the displacement direction of the tower wind vibration sensitive node and the tension fluctuation direction of the conductor wind vibration sensitive node in the spatial coupling area, and mark the coupling node pairs whose displacement direction and tension fluctuation direction satisfy the same or opposite relationship as tower-conductor wind vibration coupling resonance node pairs.

[0177] The principal vibration direction of the displacement response time history of the tower's wind-induced vibration sensitive node is extracted using principal component analysis. The principal fluctuation direction of the tension fluctuation of the conductor's wind-induced vibration sensitive node is also extracted. The angle between these two directions is calculated. If the angle is close to 0 degrees (in the same direction) or close to 180 degrees (in opposite directions), it indicates a strong coupling relationship, and this node pair is marked as a coupled resonance node pair of the tower and conductor wind-induced vibration. All coupled resonance node pairs are output.

[0178] Step S1010: The method further includes: performing automatic micro-topographic unit segmentation processing on the original point cloud data set of the power transmission corridor; extracting ground point cloud subsets from the separated ground point cloud units in the original point cloud data set of the power transmission corridor; reconstructing the ground point cloud subsets into triangular meshes of the ground surface to generate a digital surface model of the power transmission corridor micro-topography; performing multi-scale analysis of terrain undulation on the digital surface model of the power transmission corridor micro-topography; traversing the digital surface model of the power transmission corridor micro-topography with different spatial analysis window sizes; and extracting the surface elevation undulation amplitude features under each spatial analysis window size.

[0179] A subset of the ground point cloud is extracted from the ground point cloud at P0. The Delaunay triangulation algorithm is used to reconstruct the triangular mesh of the ground surface, generating a digital land model. A series of spatial analysis window sizes, from smallest to largest, are defined. Within each window size, the digital land model is traversed, and the difference between the maximum and minimum surface elevation values ​​within the window is calculated as the undulation amplitude. The statistical characteristics of the undulation amplitude under each window size are recorded.

[0180] Step S1011: Compare the inter-scale variation trends of the surface elevation undulation amplitude characteristics under different spatial analysis window sizes, identify the scale inflection points where the surface elevation undulation amplitude changes abruptly as the spatial analysis window size increases, and determine the optimal analysis window size for terrain unit segmentation based on the scale inflection points.

[0181] Plot the curve of undulation amplitude as a function of window size. Calculate the first derivative of the curve. Find the inflection point where the first derivative abruptly changes, i.e., the absolute value of the derivative decreases and then flattens out. The window size corresponding to this inflection point is the optimal analysis window size for terrain unit segmentation. This optimal analysis window size can effectively distinguish the differences in undulation characteristics between different micro-terrain units.

[0182] Step S1012: Perform terrain unit boundary detection on the digital surface model of the power transmission corridor micro-topography using the optimal analysis window size, extract the spatial transition zone position of the surface elevation undulation amplitude, use the transition zone position as the dividing boundary line between adjacent micro-topographic units, and perform regional segmentation on the digital surface model of the power transmission corridor micro-topography along the dividing boundary line to obtain multiple independent micro-topographic units. Each independent micro-topographic unit corresponds to one of the following micro-topographic types: ridge terrain, valley terrain, hillside terrain, mountain pass terrain, or plain terrain.

[0183] The system moves pixel-by-pixel on the digital terrain model with the optimal window size. The average curvature and Gaussian curvature of the surface within the window are calculated at each pixel location. Spatial locations where curvature values ​​change positively or negatively are designated as transition zones at the boundaries of terrain units. A watershed segmentation algorithm is used to segment the region along the transition zones, resulting in multiple independent micro-topographic units. The terrain type is determined based on the morphological characteristics of the surface within each unit.

[0184] Step S1013: Extract topographic morphological feature parameters for each independent micro-topographic geomorphic unit, including the average slope direction vector, average slope angle, surface roughness, and anisotropic direction of topographic undulation for the independent micro-topographic geomorphic unit.

[0185] For each micro-topographic unit, calculate the slope vector of each triangular facet in its internal digital surface model. The average slope direction vector is taken as the direction of the average value of the slope vectors of each triangular facet. The average slope angle is taken as the arithmetic mean of the angles of each triangular facet. Surface roughness is taken as the ratio of the actual surface area to the projected area of ​​the surface. The direction of topographic relief anisotropy is taken as the direction of the maximum range of the surface elevation variability function in each direction.

[0186] Step S1014: Spatial registration is performed between the topographic morphological feature parameters of each independent micro-topographic unit and the spatial trajectory parameter equation of the catenary of the traverse component in the corresponding spatial segment, to determine the type of micro-topographic unit and the corresponding topographic morphological feature parameters traversed by each traverse segment.

[0187] The horizontal projection of the traverse catenary is overlaid with the spatial boundaries of the micro-topographic units for analysis. The start and end arc length parameters of each micro-topographic unit traversed by the traverse are identified. The topographic type and morphological characteristic parameters of each micro-topographic unit are associated with the corresponding traverse arc length segments.

[0188] Step S1015: Perform a topographic adaptability analysis of the catenary morphology for the traverse sections that cross the boundaries of different micro-topographic units, and extract the degree of asymmetry in the sag of the catenary on both sides of the boundary of the micro-topographic unit and the degree of force imbalance at the catenary suspension point.

[0189] Take traverse sections of a certain arc length on both sides of the boundary of the micro-topographic unit. Calculate the ratio of the maximum sag of the catenary to that of the left and right sections as the degree of sag asymmetry. Calculate the difference in the horizontal tension components of the traverse on both sides of the boundary to the suspension point from the catenary parametric equation as the degree of force imbalance at the suspension point.

[0190] Step S1016: Based on the degree of asymmetry in the sag of the catenary and the degree of force imbalance at the suspension point of the catenary, and combined with the topographic morphology parameters of the micro-topographic unit, identify the asymmetrical stress sections of the conductor catenary caused by the differences in the micro-topographic unit. Link the spatial location information of the asymmetrical stress sections of the conductor catenary, the type of the micro-topographic unit, and the topographic morphology parameters and output them to generate a micro-topographic stress zoning distribution map of the transmission conductor.

[0191] Conductor arc length segments where the degree of sag asymmetry exceeds a preset asymmetry threshold or the degree of stress imbalance exceeds a preset imbalance threshold are marked as asymmetrical stress segments of the conductor catenary. For this segment, its start and end arc length parameters, start and end three-dimensional coordinates, the type of micro-topographic units on both sides, and the topographic morphological characteristic parameters of each topographic unit are output. The information of all asymmetrical stress segments is integrated to generate a micro-topographic stress zoning distribution map of the transmission conductor.

[0192] In an exemplary embodiment, a three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space is provided. This system can be a terminal, server, etc., and includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, near-field communication, or other technologies. When the computer program is executed by the processor, it implements a method for three-dimensional feature processing of power transmission lines based on micro-topographic point cloud space. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device can be a touch layer covering the display screen, or a button, trackball, or touchpad set on the shell of the three-dimensional feature processing system for power transmission lines based on micro-topography point cloud space, or an external keyboard, touchpad, or mouse, etc.

[0193] The following is combined Figures 3-5 This paper introduces the simulation process of a three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space, such as... Figure 3As shown, this embodiment demonstrates the loading and preliminary processing of the raw point cloud data set of the power transmission corridor collected by the UAV-borne LiDAR. This includes a 3D perspective view of the raw point cloud data set and the corresponding elevation histogram adaptive ground filtering processing results. View A uses a 3D perspective (X, Y, Z) to display a massive discrete point cloud scatter plot. The point cloud color mapping transitions from dark gray to light gray based on the "echo intensity reflectivity information." The X / Y axes are labeled with the "power transmission corridor" planar coordinates, and the Z axis is labeled with the "3D spatial coordinate information" elevation. Figure 3 Based on real R&D data, the total number of point clouds is 2,500,000 points, with coordinate ranges X[0, 200m], Y[0, 50m], and Z[0, 30m]. The echo intensity in high point cloud density areas (tower base) is >0.8, while the echo intensity in low point cloud density areas (vegetation) is <0.3. Layout B consists of an upper point cloud view and a lower histogram. In the upper point cloud view, the point cloud is color-coded with green dots labeled "ground point cloud" and red dots labeled "non-ground point cloud." Green dots are erased, leaving only red dots labeled "generated overhead point cloud set." The lower histogram has X-axis labeled "elevation interval" (meters) and Y-axis labeled "point cloud frequency distribution." The graph is a bar chart, with the vertical line labeled "segmentation threshold" (Z=2.5m). B is based on real R&D data. Peak 1 (ground) is located at an elevation of 0-3m with a frequency of 1,200,000. Peak 2 (above ground) is located at an elevation of 10-25m with a frequency of 1,300,000. The segmentation threshold of 2.5m is located at the bottom of the valley between the two peaks.

[0194] like Figure 4 As shown in the figure, this embodiment demonstrates the refined processing flow of an overhead point cloud set. C represents the result of voxelized spatial connectivity clustering, and D represents the result of principal component analysis axial feature extraction. Figure 4A two-dimensional top-down raster map (X, Y) is used, with the image being a grid-filled map. Grid cells are labeled "regular voxel grid," and the dimensions are labeled "preset voxel size 0.5m x 0.5m x 0.5m." The grid lines are thin black lines labeled "determining the continuity of point cloud density between adjacent voxels." Blue connected regions in the filled color blocks are labeled "tower components," and yellow connected regions are labeled "conductor components." C) Based on real R&D data, the clustering results show that component 1 (tower) occupies 150 voxels at position X=50m, component 2 (tower) occupies 150 voxels at position X=100m, and component 3 (conductor) is a long strip-shaped connected region with a length of 200m and a width of 2m. Isolated noise points are filtered out and not displayed. Layout D consists of a 3D point cloud on the left and a parameter table on the right. The 3D point cloud on the left only displays the point cloud of the "conductor component," with a thick red line superimposed to indicate the "main extension direction." The formula is labeled "the eigenvector corresponding to the minimum eigenvalue." The auxiliary line is the horizontal reference plane, and the angle is labeled "2.5° with the horizontal plane." The parameter table on the right has the headers "Independent Connected Components," "Eigenvalues ​​λ1, λ2, λ3," "Main Extension Direction," and "Angle with the Horizontal Plane." The data rows show component A (conductor): λ=[0.01, 0.05, 0.94], angle 2.5°, and component B (tower): λ=[0.85, 0.10, 0.05], angle 88.0°. The decision box next to component A is marked in red with "Less than the preset acute angle threshold (5°) is marked as a conductor component." Based on real R&D data, the minimum feature value of the conductor component is λ_min=0.01, the direction vector is [0.999, 0.001, 0.044], and the judgment threshold is 5°.

[0195] like Figure 5 As shown in the figure, this embodiment illustrates the three-dimensional spatial trajectory parameterization process of the conductor assembly. Figure 5 A three-dimensional curve (X, Y, Z) is used, which is a superposition of discrete points and fitted curves. The blue scattered points are labeled as "conductor assembly" discrete point cloud, the red smooth curve is labeled as "cubic spline interpolation space curve fitting processing result", and the curve equation is labeled as "conductor catenary parametric equation". Figure 5 It includes a parameter output box containing "three-dimensional spatial trajectory parameters of the conductor". The parameter list includes span: 200m, sag: 3.5m, and suspension point height difference: 1.2m. Figure 5 Based on real R&D data, the fitting residual RMSE < 0.05m, with 5 control points (per level), and output parameters A = 1200, B = 0.02, C = 15.3 (catenary equation coefficients).

[0196] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for processing three-dimensional features of power transmission lines based on micro-topographic point cloud space, characterized in that, The method includes: Load the original point cloud data set of the power transmission corridor collected by the UAV airborne lidar. The original point cloud data set of the power transmission corridor includes the three-dimensional spatial coordinate information and echo intensity reflectivity information of each scanning point. The original point cloud data set of the power transmission corridor is subjected to elevation histogram adaptive ground filtering processing. The point cloud frequency distribution in each elevation interval is statistically analyzed. The segmentation threshold between ground points and non-ground points is determined. Ground point clouds are removed to generate an overhead point cloud set, which includes tower point clouds and conductor point clouds. The overhead point cloud set is subjected to voxelized spatial connectivity clustering processing. The point cloud space is divided into regular voxel grids with a preset voxel size. The continuity of point cloud density between adjacent voxels is determined. The overhead point cloud set is clustered into multiple independent connected components. The multiple independent connected components include tower components and conductor components. Principal component analysis is performed on each of the independent connected components to extract axial features. The feature vector corresponding to the minimum feature value of the point cloud of each component is calculated as the main extension direction. Components whose main extension direction has an angle with the horizontal plane that is less than a preset acute angle threshold are marked as wire components. The conductor assembly is subjected to cubic spline interpolation spatial curve fitting processing to fit the discrete point cloud into a smooth catenary parametric equation, and the three-dimensional spatial trajectory parameters of the conductor are output.

2. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The process of performing elevation histogram adaptive ground filtering on the original point cloud data set of the power transmission corridor, statistically analyzing the point cloud frequency distribution within each elevation interval, determining the segmentation threshold between ground points and non-ground points, removing ground point clouds, and generating an overhead point cloud set includes: The elevation direction coordinate components of each scanning point are extracted from the original point cloud data set of the power transmission corridor to generate an elevation value sequence of all scanning points. The maximum and minimum elevation values ​​in the elevation value sequence of all scanning points are searched to construct the total elevation distribution interval. The total elevation distribution interval is evenly divided into multiple elevation statistical intervals. The elevation value sequence of all scanning points is traversed, and each scanning point is assigned to the corresponding elevation statistical interval according to its elevation value. The cumulative frequency of scanning points falling into each elevation statistical interval is counted to generate an elevation interval point cloud frequency distribution sequence. The frequency difference operation between adjacent intervals is performed on the frequency distribution sequence of the point cloud in the elevation interval to obtain the frequency change gradient between each adjacent elevation statistical interval. The boundary of the interval with the largest absolute value of the frequency change gradient is determined as the initial segmentation elevation threshold between the ground point and the overhead point. A local elevation neighborhood window is constructed with the initial segmentation elevation threshold as the center. The scan points in the local elevation neighborhood window are extracted to form a neighborhood point set. The elevation value standard deviation of the neighborhood point set is calculated and compared with the preset elevation fluctuation tolerance value. When the standard deviation of the elevation value is less than the preset elevation fluctuation tolerance value, the initial selected segmentation elevation threshold is determined as the final segmentation elevation threshold. The final segmentation elevation threshold is used to perform binary segmentation on the original point cloud data set of the power transmission corridor. Scanning points with elevation values ​​less than the final segmentation elevation threshold are classified as ground point clouds, and scanning points with elevation values ​​greater than or equal to the final segmentation elevation threshold are classified as overhead point clouds. Remove all scan points classified as ground point clouds from the original point cloud data set of the power transmission corridor, and retain all scan points classified as overhead point clouds to form an overhead point cloud set. The overhead point cloud set contains the three-dimensional spatial coordinate information and echo intensity reflectivity information of each overhead scan point. The spatial distribution of the overhead point cloud set is statistically analyzed. The overhead point cloud set is divided into multiple elevation layers according to the elevation direction. The number of overhead scanning points in each elevation layer is counted, and an overhead point cloud elevation layer distribution histogram is generated. Based on the trend of point cloud quantity change in each elevation layer in the overhead point cloud elevation layer distribution histogram, the clustered and sparse segments of point cloud quantity in the elevation direction are extracted. The scanning points corresponding to the clustered segments are pre-marked as potential point clouds of towers, and the scanning points corresponding to the sparse segments are pre-marked as potential point clouds of conductors. Local point cloud density calculations are performed on the potential point cloud of the tower and the potential point cloud of the conductor respectively to obtain the point cloud density values ​​of the potential area of ​​the tower and the potential area of ​​the conductor. Based on the density difference characteristics between the point cloud density values ​​of the potential area of ​​the tower and the potential area of ​​the conductor, preliminary separation marking results of the tower and conductor are generated.

3. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The process of performing voxelized spatial connectivity clustering on the overhead point cloud set involves dividing the point cloud space into a regular voxel grid with a preset voxel size, determining the continuity of point cloud density between adjacent voxels, and clustering the overhead point cloud set into multiple independent connected components, including: Extract the three-dimensional spatial coordinate information of all overhead scanning points from the overhead point cloud set, search for the minimum and maximum coordinate values ​​along the three-dimensional coordinate axis direction, construct a three-dimensional spatial bounding box that completely surrounds the overhead point cloud set, and record the span range of the three-dimensional spatial bounding box in each coordinate axis direction. Using a preset voxel side length as the basic dividing unit, the three-dimensional bounding box is uniformly divided along the three coordinate axes to generate a voxel grid network with row index number, column index number and layer index number. Each voxel grid unit in the voxel grid network corresponds to a cubic subspace region in the three-dimensional space. Traverse each overhead scanning point in the overhead point cloud set, calculate the row index number, column index number and layer index number to which the overhead scanning point belongs based on the three-dimensional spatial coordinate information of the overhead scanning point, and assign the overhead scanning point to the corresponding voxel grid cell to complete the voxel mapping of all overhead scanning points. The number of overhead scan points contained in each voxel grid cell in the voxel grid network is counted. Voxel grid cells containing a number of overhead scan points greater than zero are marked as occupied voxels, and voxel grid cells containing a number of overhead scan points equal to zero are marked as vacant voxels, generating a binary labeling map of voxel occupancy status. A neighborhood connectivity scan is performed on the binary marker map of voxel occupancy status. Each voxel grid cell marked as an occupancy voxel is traversed, and the occupancy status of the six adjacent voxel grid cells in the face-adjacent direction of the voxel grid cell is checked. The adjacent voxel grid cells that are adjacent to the occupancy voxel and are also occupancy voxels are recorded as connected adjacent voxels. Using each occupied voxel that is not assigned to any connected component as a seed voxel, a three-dimensional spatial region growth operation is performed to merge the seed voxel and all connected neighboring voxels directly connected to the seed voxel into a connected voxel cluster. The occupied voxels newly merged into the connected voxel cluster are used as new growth front voxels to continue searching for their respective connected neighboring voxels. When the voxel at the growth front of the connected voxel cluster can no longer find an unvisited connected neighboring voxel, the growth process of the connected voxel cluster is terminated. The set of overhead scanning points contained in all occupied voxels within the connected voxel cluster is extracted as an independent connected component. Component scale feature analysis is performed on the independent connected component, and the projected span length of the independent connected component in each coordinate axis direction in three-dimensional space is calculated. The spatial morphological extension of the independent connected component is determined according to the ratio of the maximum span direction to the minimum span direction in the projected span length. Independent connected components with spatial morphological extension greater than a preset extension threshold are initially classified as conductor components, and independent connected components with spatial morphological extension less than or equal to the preset extension threshold are initially classified as pole components, generating preliminary classification results for independent connected components. A cross-component spatial adjacency analysis is performed on the preliminary classification results of the independent connected components. The number of adjacency faces between the independent connected components initially classified as conductor components and the independent connected components initially classified as pole components in the voxel grid network is extracted. The classification label of the independent connected components with adjacent boundaries is corrected based on the number of adjacency faces.

4. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The step of performing principal component analysis axial feature extraction on each of the independent connected components, calculating the feature vector corresponding to the minimum feature value of each component's point cloud as the main extension direction, and marking components whose main extension direction makes an angle less than a preset acute angle threshold with the horizontal plane as wire components includes: For each independent connected component, the three-dimensional spatial coordinate information of all scan points within the independent connected component is extracted, a local coordinate system is constructed with the centroid of the independent connected component as the origin, the three-dimensional spatial coordinate information of all scan points is transformed into the local coordinate system to obtain a centroid-free spatial coordinate set, and the point cloud spatial scatter matrix of the independent connected component is constructed using the centroid-free spatial coordinate set. Each element of the point cloud spatial scatter matrix is ​​the coordinate component covariance of all scan points within the independent connected component in the corresponding coordinate axis direction. The point cloud spatial scattering matrix is ​​subjected to eigenvalue decomposition to solve for all eigenvalues ​​of the point cloud spatial scattering matrix and the eigenvector corresponding to each eigenvalue. The solved eigenvalues ​​are arranged in descending order of numerical value to generate a descending sequence of eigenvalues. The eigenvectors are arranged in the order of their corresponding eigenvalues ​​to form a sequence of eigenvectors. The smallest eigenvalue is extracted from the descending sequence of eigenvalues. The smallest eigenvector corresponding to the smallest eigenvalue is extracted from the sequence of eigenvectors. The direction pointed to by the smallest eigenvector is determined as the minimum spatial scattering direction of the independent connected component. The minimum spatial dispersion direction is taken as the main extension direction of the independent connected component, and the main extension direction represents the narrowest dimensional direction of the point cloud spatial distribution within the independent connected component. Construct a reference horizontal plane normal vector, the direction of which is perpendicular to the reference horizontal plane and points in the direction of increasing elevation; calculate the angle between the main extension direction and the reference horizontal plane normal vector. The directional angle is compared with a preset acute angle threshold. When the directional angle is less than the preset acute angle threshold, the independent connected component is marked as a wire component. When the directional angle is greater than or equal to the preset acute angle threshold, the independent connected component is marked as a non-wire component. For independent connected components marked as wire assemblies, the consistency of the main extension direction is verified. The projected coordinate components of each scanning point in the wire assembly along the main extension direction are extracted, and the distribution span length and distribution center position of the projected coordinate components are calculated. Based on the distribution span length of the projected coordinate components and the distribution center position, combined with the cross-sectional point cloud distribution range of the conductor assembly on the plane perpendicular to the main extension direction, a spatial extension direction envelope of the conductor assembly is constructed. The morphological regularity of the spatial extension direction envelope is analyzed, and the cross-sectional area change curve of the spatial extension direction envelope along the main extension direction is calculated. The fluctuation amplitude parameter of the cross-sectional area change curve is extracted, and the conductor type confidence of the conductor assembly is adjusted according to the fluctuation amplitude parameter.

5. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The process of performing cubic spline interpolation spatial curve fitting on the conductor assembly to fit the discrete point cloud into a smooth catenary parametric equation and outputting the three-dimensional spatial trajectory parameters of the conductor includes: For the conductor assembly, the three-dimensional spatial coordinate information of all conductor scanning points within the conductor assembly is extracted. The all conductor scanning points are projected and sorted along the main extension direction of the conductor assembly to obtain an ordered sequence of conductor scanning points arranged in spatial order. The first and last conductor scanning points are extracted from the ordered sequence of conductor scanning points. The first conductor scanning point is used as the curve fitting start point and the last conductor scanning point is used as the curve fitting end point. The total arc length parameter interval of the conductor spatial extension is constructed between the curve fitting start point and the curve fitting end point. The total arc length parameter interval is evenly divided into multiple arc length parameter segment intervals, and multiple consecutive conductor scanning points are selected in each arc length parameter segment interval as the set of fitting control points for that arc length parameter segment interval. For each arc length parameter segmented interval, a cubic spline interpolation basis function is constructed using the set of fitting control points within that arc length parameter segmented interval. The cubic spline interpolation basis function satisfies the function value continuity condition, the first derivative continuity condition, and the second derivative continuity condition at the segmented connection point of adjacent arc length parameter segmented intervals. By simultaneously establishing the cubic spline interpolation basis functions for all segmented intervals of the arc length parameters, a global cubic spline interpolation curve equation system for the conductor assembly is generated. The global cubic spline interpolation curve equation system is solved to obtain the complete cubic spline interpolation space curve of the conductor assembly. The continuous trajectory expression of the complete cubic spline interpolation space curve in three-dimensional space is converted into the form of the conductor catenary parametric equation. The conductor catenary parametric equation includes the catenary sag parameter, the catenary horizontal tension coefficient, and the spatial position parameter of the catenary suspension point. The three-dimensional spatial trajectory parameters of the conductor assembly are extracted from the catenary parametric equation of the conductor. The three-dimensional spatial trajectory parameters of the conductor include the continuous path coordinate sequence of the conductor in three-dimensional space, the spatial curvature value of the conductor at each arc length parameter position, and the spatial deflection value of the conductor at each arc length parameter position. The catenary morphology features of the three-dimensional spatial trajectory parameters of the conductor are decomposed, and the continuous path coordinate sequence in the three-dimensional spatial trajectory parameters of the conductor is decomposed into a two-dimensional planar trajectory on the horizontal projection plane and a sag variation curve in the vertical projection direction. The horizontal azimuth sequence and horizontal deflection radius sequence of the conductor assembly are extracted based on the two-dimensional plane trajectory, and the position of the maximum sag amplitude and the position of the sag symmetry center of the conductor assembly are extracted based on the sag variation curve. The horizontal azimuth sequence, the horizontal deflection radius sequence, the maximum overhang amplitude position, the overhang symmetry center position, and the spatial curvature and spatial deflection values ​​in the three-dimensional spatial trajectory parameters of the conductor are combined to generate a complete set of conductor spatial morphology description parameters, and the set of conductor spatial morphology description parameters is output.

6. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The method further includes: The tower assembly is subjected to tower head structure identification processing. A tower head point cloud subset and a tower body point cloud subset are extracted from the tower assembly. The tower head point cloud subset corresponds to the spatial point cloud of the crossarm and hanging point at the top of the tower. The tower body point cloud subset corresponds to the spatial point cloud of the main material and diagonal material of the tower. The spatial location of the hanging point is processed for the subset of tower head point cloud. Local point cloud clusters extending in the horizontal direction are searched in the subset of tower head point cloud. The spatial geometric center coordinates of each local point cloud cluster are extracted as candidate hanging point spatial locations. Calculate the horizontal spatial distance between adjacent candidate wire hanging points, verify the rationality of the horizontal spatial distance based on the standard phase-to-phase distance range in the transmission line design specification, eliminate candidate wire hanging point spatial locations that do not conform to the standard phase-to-phase distance range, and obtain the final wire hanging point spatial locations. Based on the spatial distance association matching between the final hanging point and the spatial position of the conductor endpoint of the conductor assembly, a connection correspondence is established between the conductor endpoint with the shortest spatial distance and the hanging point to determine the hanging position of each conductor assembly on the tower. The conductor arrangement sequence of the conductor assemblies associated with each splicing position on the same tower is analyzed, and the arrangement direction and spacing of each conductor assembly on the horizontal projection plane are extracted to generate conductor phase sequence arrangement layout information.

7. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The method further includes: Analysis of conductor morphology changes within the span of conductor assemblies between two adjacent towers is performed. Spatial position sampling points of conductors are extracted at equal arc length intervals along the spatial extension direction of the conductors. The spatial distance increment and directional change angle between adjacent conductor spatial position sampling points are calculated. The spatial distance increment and the directional change angle are input into the catenary tension state inversion model to invert and calculate the real-time horizontal tension distribution of the conductor assembly within the span, thereby obtaining the tension change curve along the conductor extension direction. Based on the tension change curve, the tension peak position and tension valley position are extracted. The spatial correspondence between the tension peak position and the spatial coordinates of the terrain protrusion point in the micro-topography point cloud data is analyzed to identify the local influence area of ​​the terrain protrusion on the conductor tension. Safety clearance verification is performed on conductor assemblies within the local impact area. The vertical distance between the conductor spatial location sampling point and the point cloud of the terrain surface below is calculated. The vertical distance is compared with the minimum safety distance specified in the transmission line safety regulations, and conductor safety risk sections with insufficient vertical distance are marked. By integrating the spatial location information of the conductor safety risk section, the tension value of the corresponding section in the tension change curve, and the spatial coordinates of the topographic protrusion point, a conductor tension-topographic associated risk distribution map is generated.

8. The method for processing three-dimensional features of transmission lines based on micro-topographic point cloud space according to claim 1, characterized in that, The method further includes: The conductor assembly is subjected to wind vibration sensitive section identification processing, the local bending curvature change drastic section of the conductor assembly under the catenary morphology is extracted, and the curvature change gradient amplitude of each local bending curvature change drastic section is calculated; The curvature change gradient magnitude is compared with a preset curvature change gradient threshold. The segment where the curvature change gradient magnitude is greater than the preset curvature change gradient threshold is marked as a morphological change sensitive segment. The morphological change sensitive segment corresponds to the position where the degree of curvature changes drastically in the catenary morphology of the conductor. The conductor cross-section orientation analysis is performed on the morphologically sensitive section, the cross-section normal vector of the conductor in the morphologically sensitive section is extracted at each sampling position, and the incident angle between the cross-section normal vector and the prevailing wind direction is calculated. If the incident angle exceeds a preset angle threshold and the magnitude of the curvature change gradient exceeds a preset curvature change gradient threshold, then the sensitive section of the morphological change is identified as the target risk section. Local point cloud density enhancement sampling is performed on the point cloud of the conductor within the target risk section, and the target risk section is repeatedly scanned from multiple angles by the UAV-borne lidar to obtain supplementary point cloud data for the target risk section.

9. A three-dimensional feature processing system for power transmission lines based on micro-topographic point cloud space, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the three-dimensional feature processing method for power transmission lines based on micro-topographic point cloud space as described in any one of claims 1 to 8 by executing the machine-executable instructions.

10. A machine-readable storage medium for storing machine-executable instructions of a processor; in, The processor is configured to execute the method for three-dimensional feature processing of power transmission lines based on micro-topographic point cloud space as described in any one of claims 1 to 8 by executing the machine-executable instructions.

Images