Method and system for processing radar scanning point cloud data of power corridor line

Abstract

The invention relates to the technical field of automatic inspection of power corridor lines, and discloses a method and a system for processing radar scanning point cloud data of a power corridor line. According to the method, an unsupervised point cloud segmentation system and an advanced three-dimensional data processing technology are applied, point cloud data collected by an unmanned aerial vehicle are deeply analyzed, key elements of a power transmission line and the surrounding environment of the power transmission line are automatically extracted, the key elements comprise electric wires, electric towers, buildings, roads, water bodies, vegetation and the like, and accurate semantic segmentation is performed; therefore, potential safety hazards are effectively identified and early warned, and efficient and automatic routing inspection of the electric power corridor line is realized. The method effectively overcomes the defects of low efficiency and low accuracy of a traditional manual inspection mode, and ensures that the safety and economical efficiency of power grid operation are remarkably improved. According to the method, a scientific data driving method is applied, the accuracy and efficiency of power line inspection are improved, and powerful data support is provided for resource planning and management decision making of power enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic inspection of power corridor lines, and particularly to a method and system for processing radar scan point cloud data of power corridor lines. Background Art

[0002] As the main way of long-distance power transmission, power corridor lines are of fundamental importance in the grid infrastructure. Since power corridor lines cover complex and diverse scenarios such as vegetation, buildings, roads, and water areas, maintaining their safety has always been a key task in grid maintenance. For a long time, transmission lines could only be inspected manually, with many drawbacks, including low efficiency, low accuracy, and lack of coverage in many areas such as mountains and water areas.

[0003] To solve these problems, researchers have proposed automatic detection technologies, especially the combination of drones and lidar, which has changed the paradigm of line detection. Specifically, drones equipped with lidar can scan the transmission corridor to obtain the three-dimensional point cloud of the transmission line and its surrounding environment, providing sufficient data for on-site analysis and hazard warning. For the processing of point cloud data, various elements such as power lines, electric towers, vegetation, buildings, and roads are first extracted. Then, the distances between the power lines and the detected elements can be calculated, and potential hazard points can be found to prevent further safety risks.

[0004] In the above process, identifying objects in point cloud data is a key link. At present, many scholars have made a series of attempts, including: (1) Power line detection: A method that assumes that the shape of power lines is similar to that of parabolas is adopted, and the point cloud data of power lines is fitted to the parabola model through least squares regression. In addition, a robust estimator is introduced to improve the accuracy of extracting power lines from LiDAR point cloud data. Another method is to identify power lines based on the shape characteristics of power lines using specific evaluation indicators. There is also a method of extracting power lines with linear features by constructing voxel grids. (2) Power tower detection: A technology that combines image and point cloud data is used to detect power towers. This method first locates the position of the tower in a two-dimensional image and then accurately extracts it in a three-dimensional point cloud. Another technology determines the position of the power tower by three-dimensional grid division and spatial line fitting. Another method uses height gaps to define a robust seed region and extracts the complete power tower structure through a region growing algorithm. (3) Building detection: A method uses flatness, normal variance, and uniformity of gray-level co-occurrence matrix (GLCM) as clues to identify buildings. Another method uses a clustering algorithm based on Euclidean distance to identify buildings. There is also a method that uses I-Octree technology to automatically identify the roof points of buildings. In addition, a scan line method based on Euclidean distance is proposed to extract buildings. (4) Road detection: One method identifies road points with unimodal characteristics by calculating the histogram of normal vectors. Another method combines local road features and global attributes to extract roads. There is also a ground segmentation algorithm based on lidar histogram for fast segmentation of roads in urban environments. At the same time, some researchers have adopted a triangular irregular network strategy to detect water areas.

[0005] Although existing technologies have alleviated the difficulties in power corridor detection to a certain extent, there is still room for improvement in data processing speed, accuracy and adaptability to complex environments. Summary of the invention

[0006] The technical problem to be solved by the present invention is to propose a method and system for processing radar scanning point cloud data of power corridor lines, so as to achieve more efficient and accurate automatic detection of power corridor lines that can handle complex geographical environments.

[0007] In order to solve the above technical problems, the present invention first proposes a method for processing laser radar scanning point cloud data of a power corridor line, comprising the following steps:

[0008] a. Data preprocessing: mainly includes uniform downsampling and through filtering of point clouds;

[0009] b. Point cloud data extraction of power lines;

[0010] c. Extraction of point cloud data of the tower;

[0011] d. Point cloud data extraction of buildings: Calculate the angle of the normal vector of any point cloud data. If the angle of the normal vector meets certain conditions, it is considered that the point cloud data is building point cloud; then analyze the connected regions in the building point cloud and filter out the regions smaller than the specified area threshold according to the area of each connected domain;

[0012] e. Point cloud data extraction of roads: Use the RANSAC plane fitting algorithm to fit the ground points onto a plane; calculate the Euclidean distance of the color between any given point and the road to obtain the similarity between the given point and the road, and finally screen out the road point cloud data according to the similarity value;

[0013] f. Point cloud data filling of water systems: Project the ground point cloud onto the X and Y planes, and use the interpolation method to fill the holes in the point cloud, which is the point cloud data of the water system.

[0014] Preferably, the specific method of filling the point cloud data of the water system is as follows:

[0015] f-1. Perform Delaunay triangulation on the X and Y coordinates of the point cloud data to obtain the triangular mesh of the point set;

[0016] f-2. By traversing each triangle of the triangular mesh, extract its edges and record the number of occurrences of each edge. The edges that appear once are considered boundary edges;

[0017] f-3. According to the boundary edges, extract the boundary vertices, and then generate a convex hull through the Shapely library to obtain a smooth boundary contour;

[0018] f-4. Create a grid in the contour, generate grid lines and check whether the center points of the grids are inside the shrunk contour. If inside, mark them as valid grid points;

[0019] f-5. Perform a quick query on the valid grids to determine whether there is point cloud inside the grids; if there is no point cloud, fill it with an empirical value of points;

[0020] f-6. Perform DBSCAN clustering on the filled point cloud, and screen according to the area of each cluster, and retain the clusters with an area greater than a certain threshold;

[0021] f-7. Reassign the Z coordinates of the cluster point cloud obtained in step f-6, that is, obtain the point cloud data of the water system with three-dimensional coordinates.

[0022] The connected domain analysis method in the building is based on the determination rule of adjacent pixels, that is, the area of a single pixel pixel size The calculation formula is:

[0023] In the formula, x min and x max is the minimum and maximum value of the x coordinate after projection, and y min and max The minimum and maximum values ​​of the projected y coordinate; image_resolution is the resolution of the image, that is, the number of grids.

[0024] The calculation formula of the area A of the connected domain is: A = n·pixel size , where n is the number of buildings.

[0025] The calculation formula for the area of ​​each cluster in step f-6 is:

[0026] In the formula, S represents the area of ​​the polygon; n represents the number of polygon vertices, that is, the number of cluster boundary points; (x i ,y i ) is the coordinate of the ith vertex, (x i+1 ,y i+1 ) is the coordinate of the i+1th vertex;

[0027] The method for reassigning the Z coordinate of the cluster point cloud obtained in step f-6 is:

[0028] Find the nearest neighbor of the sampling point in the original point cloud and calculate the average value of its Z coordinate according to the following formula: new And assign it to all points in the cluster, In the formula, z new is the average value of the new z coordinates in the cluster, N represents the number of points in the cluster, and z i is the z coordinate of the i-th point in the nearest neighbor.

[0029] The present invention also provides an application system with the above point cloud data processing method as the core, which is convenient for terminal users to use. The user first initiates a request to the back end through the front end page, and the request is received and processed by the high-performance HTTP server interface;

[0030] The high-performance HTTP server interface interacts with the point cloud database obtained after point cloud data processing, and queries the directory and path information of the required files through the point cloud database;

[0031] After the point cloud database returns the query result, the high-performance HTTP server interface obtains the file path or actual file content and returns it to the front-end page, or the high-performance HTTP server interface obtains the actual file through the file path and generates a report file;

[0032] End users download point cloud files or report files through the page.

[0033] Furthermore, during the file transfer process, multi-level monitoring is used, including

[0034] Level 1 monitoring: Detect that both the las point cloud file and the kml file are transferred to the specified folder, and start the code to run;

[0035] Level 2 monitoring: After the backend code ends, generate las files with different segmentation results in the specified folder, start the "Start potreeconverter" conversion tool, and perform point cloud format conversion;

[0036] Level 3 monitoring: Monitor the folder where the conversion is completed. After a point cloud file comes in, store the information corresponding to the point cloud file in the database.

[0037] The present invention uses an unsupervised point cloud segmentation system and advanced 3D data processing technology. By deeply analyzing the point cloud data collected by drones, key elements of the transmission line and its surrounding environment, including wires, electric towers, buildings, roads, water bodies, and vegetation, etc., are automatically extracted, and precise semantic segmentation is performed, thereby effectively identifying and warning potential safety hazards, and realizing efficient and automated inspection of power corridor lines in complex geographical environments. The present invention uses a scientific data-driven method, which not only improves the accuracy and efficiency of power line inspection, but also provides strong data support for the resource planning and management decision-making of power enterprises. The present invention effectively overcomes the low efficiency and low accuracy of the traditional manual inspection method, greatly improves the monitoring efficiency of power corridor lines, enhances safety, reduces maintenance costs, and ensures that the safety and economy of power grid operation are significantly improved, and has important technical significance and application prospects. Brief Description of the Drawings

[0038] The technical solutions of the present invention will be further specifically described below in conjunction with the drawings and specific embodiments.

[0039] Figure 1 It is the overall implementation flowchart of the present invention.

[0040] Figure 2 It is the DBSCAN clustering algorithm process adopted for the extraction of power tower pole point cloud data.

[0041] Figure 3 It is the RANSAC algorithm process adopted for the extraction of ground point cloud data.

[0042] Figure 4 It is the application system framework diagram of the present invention.

[0043] Figure 5 It is the specific implementation style of the report file.

[0044] Figure 6 It is a screenshot of the user visualization front-end page. Detailed implementation manners

[0045] The automatic detection method for power corridor lines proposed by the present invention, which is more efficient, accurate and capable of handling complex geographical environments, includes the following steps:

[0046] 1. Data preprocessing: Due to actual factors in scanning, the point cloud is affected by isolated points, flying points and noise, which affects the accurate extraction of elements. Therefore, preprocessing is required. The preprocessing mainly includes uniform downsampling and filtering of the point cloud. Uniform downsampling reduces the data volume and computational complexity. The pass-through filter removes noise points and outliers.

[0047] 2. Extraction of power line point cloud data: The power line is the most important element in the task. The power line is detected through evaluation and linearity. By analyzing the elevation distribution of the point cloud, it can be noted that the power line has the characteristics of higher elevation and discontinuous elevation. Therefore, the elevation is used to remove the point cloud close to the ground, and only the power line, some high vegetation and tower points are retained. After that, since the power line is linearly distributed while other points are irregularly scattered, the line is further extracted in a linear manner. To describe the linearity, the principal component analysis (PCA) method is used as follows:

[0048] First, select the i-th point as point P i , and the three eigenvalues λ i1 , λ i2 and λ i3 of the point set covariance matrix in its neighborhood with a radius of R can be used for a rough judgment. If the eigenvalues satisfy , then the point presents a planar feature, such as the surface of an electric tower; if the eigenvalues satisfy , then the distribution of the point is disordered and it is very likely to be a vegetation point; if the eigenvalues satisfy , it indicates that the point shows a linear distribution, which is very likely to be a power line point or a boundary point of a certain surface. Then, through the linear eigenvalue formula, the linear eigenvalues of the points whose eigenvalues satisfy are calculated. Finally, the calculated linear eigenvalues are compared with a set threshold, and the points with linear eigenvalues greater than the threshold are extracted to complete the accurate extraction of the power line. The formula is as follows, where L i is the linear eigenvalue of point P i .

[0049]

[0050] 3. Extraction of tower point cloud data: First, apply cloth filtering to divide the points into ground points and non-ground points, and only non-ground points are considered when extracting the tower.

[0051] Then, in order to quickly locate and extract the tower point cloud in a large range, as shown in formula The non-ground points are spatially divided into M×N×1 regular blocks, and it is checked whether there is a tower inside each grid, where x min and x max are respectively the minimum and maximum values in the x coordinate. y min and y max are similar in the y dimension, and s is the grid size.

[0052] Different from other elements, the elevation of the electric tower is extremely high and continuous, and the grid of the electric tower can be determined according to this feature. In order to eliminate the influence of the terrain, as shown in the formula where z(m,n,j) is the z coordinate of the j-th point in the grid of the m-th row and n-th column. Using the Digital Elevation Model (DEM, the value of which can be V DEM ) and the Digital Surface Model (DSM, the value of which can be V DSM ), the Normalized Digital Surface Model (nDSM, the value of which can be V nDSM ) is calculated to reflect the maximum height information inside the grid relative to the ground. According to the obtained nDSM, a suitable height difference threshold is set to obtain the point cloud grid with a height difference greater than the height difference threshold H;

[0053] After extracting the point cloud grid, in addition to the electric tower point cloud, the obtained point cloud also contains noise point clouds such as low vegetation around the tower base and insulators connected to the tower body. Therefore, in order to obtain a more refined high-voltage electric tower point cloud, the DBSCAN clustering algorithm needs to be used to remove the noise points.

[0054] 4. Extraction of building point cloud data: The formula is used to calculate the normal vector i of any point P and the angle with the z-axis [0,0,1] to extract the building point cloud. However, it also contains some vegetation point clouds and ground point clouds, so it is necessary to further find the connected domain in the current point cloud; project the point cloud points along the Z axis onto the XY plane, convert the projected point cloud data (XY coordinates) into a two-dimensional image, calculate the two-dimensional histogram, apply binary processing to the generated two-dimensional image, and divide the points into two categories: the area with points and the area without points. Mark the connected regions in the binary image, calculate the pixel area of each connected domain, and filter out the connected domains with an area smaller than the specified area threshold.

[0055] 5. Extraction of road point cloud data: Use the Random Sample Consensus (RANSAC) plane fitting algorithm to fit the ground points to a plane. Define the color of the road point cloud as (R d , G d , B d ), where R drepresents the red component in the color channel of the road point cloud, G d represents the green component in the color channel of the road point cloud, B d represents the blue component in the color channel of the road point cloud; through the formula calculate the color Euclidean distance between any point p(R p , G p , B p ) in the current point cloud and the road, to evaluate the similarity between any point p(R p , G p , B p ) in the current point cloud and the road. In the formula, C d represents the color Euclidean distance between the color of the road point cloud and the color of any point p in the current point cloud. (R p , G p , B p ) where R P represents the red component in the color channel of any point in the current point cloud, G P represents the green component in the color channel of any point in the current point cloud, B P represents the blue component in the color channel of any point in the current point cloud. When C d is less than a certain empirical value such as 0.4, the point cloud is regarded as a road point cloud.

[0056] 6. Filling of water system point cloud data, including the steps:

[0057] 6-1. Perform Delaunay (Delaunay Triangulation) triangulation on the X and Y coordinates of the point cloud data to obtain the triangular mesh of the point set;

[0058] 6-2. By traversing each triangle of the triangular mesh, extract its edges and record the occurrence times of each edge. The edge that appears once is considered a boundary edge;

[0059] 6-3. According to the boundary edges, extract the boundary vertices, and then generate a convex hull through the Shapely library to obtain a smooth boundary contour; Shapely is a python library for the operation and analysis of geometric objects based on Cartesian coordinates. The full English name of the Shapely library is "Shapely: Manipulation and analysis of geometric objects in the Cartesian plane."

[0060] 6-4. Create a grid in the contour, generate grid lines and check whether the center point of the grid is inside the reduced contour; if it is inside, it is marked as a valid grid point;

[0061] 6-5. Quickly query the valid grid to determine whether there is point cloud in the grid; if there is no point cloud, fill it with the number of points equal to the empirical value. In this specific embodiment, the empirical value is set to 1000;

[0062] 6-6. Perform DBSCAN ("Density-Based Spatial Clustering of Applications with Noise", the Chinese full name is "Density-Based Spatial Clustering Application with Noise") clustering on the filled point cloud, and screen according to the area of each cluster, and retain the clusters with an area greater than a certain threshold. In this specific embodiment, the aforementioned certain threshold is set to 150 square meters;

[0063] 6-7. Reassign the Z coordinates of the clustered point cloud obtained in step 6-6, that is, obtain the water system point cloud data with three-dimensional coordinates.

[0064] The connected domain analysis method in the building is based on the determination rule of adjacent pixels, that is, the area of a single pixel pixel size The calculation formula is:

[0065] In the formula, x min and x max are the minimum and maximum values of the projected x coordinates, and y min and y max are the minimum and maximum values of the projected y coordinates;;

[0066] image_resolution is the resolution of the image, that is, the number of grids.

[0067] The calculation formula for the area A of the connected domain is: A = n·pixel size , where n is the number of buildings.

[0068] The calculation formula for the area of each cluster in step f-6 is:

[0069] In the formula, S represents the area of the polygon; n represents the number of vertices of the polygon, that is, the number of cluster boundary points; (x i ,y i ) is the coordinate of the i-th vertex, and (x i+1 ,y i+1 ) is the coordinate of the (i + 1)-th vertex;

[0070] Since the ground point cloud was projected onto the X and Y planes at the beginning, the ground point cloud has lost its original elevation information at this time, that is, the Z coordinates of all points have become 0. According to empirical values, the clustered point clouds with an area greater than a certain area threshold are retained, that is, the water system point clouds. For example, the certain area threshold is 150 square meters. Since the Z coordinates of the water system point clouds are all 0 at this time, the Z coordinates of the clustered point clouds must be restored. The method for restoring the Z coordinates of the clustered point clouds described in step f-6 is:

[0071] Find the nearest neighbor of the sampling point in the original point cloud and calculate the average value of its Z coordinate according to the following formula: new And assign it to all points in the cluster, In the formula, z new is the average value of the new z coordinates in the cluster, N represents the number of points in the cluster, and z i is the z coordinate of the i-th point among the nearest neighbor points. That is, the average value of the z coordinates of all the nearest points is calculated and assigned to the current cluster point cloud.

[0072] 7. Extraction of vegetation point cloud data: After the above steps, the point clouds of power lines, towers, buildings, roads and water systems have been extracted. For further analysis, these separated point clouds are removed from the original point cloud, and the remaining point clouds are the vegetation point clouds. This step ensures that the vegetation area can be accurately identified and processed, and other types of point clouds are excluded.

[0073] In order to facilitate the implementation and application of the above processing method, the present invention also provides an application system with the above point cloud data processing method as the core for easy use by terminal users. The system framework and its interactive process are as follows: Figure 1 As shown, including:

[0074] The user first initiates a request to the backend through the front-end page, and the request is received and processed by FastAPI, which is a web framework in Python and is mainly used to write interface codes. Of course, those skilled in the art clearly know that other high-performance HTTP server interfaces similar to FastAPI can also achieve the same technical effects as FastAPI adopted in this specific implementation. Such other similar technical means are equivalent technical means and also fall within the scope of protection of the present invention;

[0075] Next, FastAPI interacts with the point cloud database obtained after point cloud data processing, and queries the directory and path information of the required files through the point cloud database;

[0076] After the point cloud database returns the query results, FastAPI obtains the file path or actual file content and returns it to the front-end page, or FastAPI obtains the actual file through the file path and generates a report file;

[0077] Finally, the end user downloads the point cloud file or report file through the page.

[0078] During the above file transfer process, multi-level monitoring is used, including

[0079] First-level monitoring: Detect that both the las point cloud file and the kml file are transferred into the specified folder, and start the code to run.

[0080] Second-level monitoring: After the backend code ends, generate las files with different segmentation results in the specified folder, start the "Start potreeconverter" conversion tool, and perform point cloud format conversion. The aforementioned "Start potreeconverter" is a conversion tool that converts 3D point cloud data in las format to an efficient WebGL format.

[0081] Third-level monitoring: Monitor the folder where the conversion is completed. After the point cloud file comes in, store the information corresponding to the point cloud file in the database.

[0082] Among them, the report generation method is as follows: According to the generated categories, use the docx library to create a table in the Word document to store the clustering results and related data, and use the open3d library to generate the three-view images of the point cloud and insert them into the Word document. The report includes the tower interval where each category is located, the distance from the small-numbered tower, the longitude and latitude, as well as the horizontal distance, vertical distance, and clearance distance from the power line, as Figure 5 shown.

[0083] The front-end page is a visualization page that displays the user interface of an unsupervised point cloud segmentation system for power corridor line patrol, mainly covering the file directory tree, point cloud data visualization area, layer control panel, measurement tool, parameter setting, report generation button, and navigation control module, as Figure 6 shown.

[0084] Finally, it should be noted that the above specific implementation manners are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for processing radar scanning point cloud data of power corridor lines, characterized in that: The following steps are involved: a. Data preprocessing: mainly includes uniform downsampling and through filtering of point clouds; b. Point cloud data extraction of power lines; c. Extraction of point cloud data of the tower; d. Extraction of building point cloud data: Calculate the angle of the normal vector of any point cloud data. If the angle of the normal vector meets certain conditions, the point cloud data is considered to be a building point cloud; then analyze the connected domains in the building point cloud, and filter out areas smaller than the specified area threshold according to the area of ​​each connected domain; e. Extraction of road point cloud data: Use the RANSAC plane fitting algorithm to fit ground points onto a plane; Calculate the Euclidean distance of the color between any given point and the road to obtain the similarity between the given point and the road, and finally filter out the road point cloud data according to the similarity value; f. Filling in point cloud data of the water system: Project the ground point cloud onto the X, Y plane, and use the interpolation method to fill in the holes in the point cloud to obtain the point cloud data of the water system.

2. The method for processing radar scanning point cloud data of power corridor lines according to claim 1 is characterized in that: The specific method of filling the point cloud data of the water system is: f-1. Perform Delaunay triangulation on the X and Y coordinates of the point cloud data to obtain a triangular mesh of the point set; f-2. By traversing each triangle of the triangular mesh, extracting its edges, and recording the number of occurrences of each edge, the edge that appears once is considered a boundary edge; f-3. Extract boundary vertices based on boundary edges, and then generate convex hulls through the Shapely library to obtain a smooth boundary contour; f-4. Create a grid in the contour, generate grid lines and check whether the center point of the grid is inside the reduced contour; if so, mark it as a valid grid point; f-5. Quickly query the valid grid to determine whether the grid contains point clouds; If it does not contain point clouds, fill it with experience points; f-6. Perform DBSCAN clustering on the filled point cloud, filter according to the area of ​​each cluster, and retain clusters with an area greater than a certain threshold; f-7. Reassign the Z coordinate of the cluster point cloud obtained in step f-6 to obtain water system point cloud data with three-dimensional coordinates.

3. The method for processing radar scanning point cloud data of power corridor lines according to claim 2 is characterized in that: The connected domain analysis method in the building is based on the judgment rule of adjacent pixels, that is, the area of ​​a single pixel size The calculation formula is: In the formula, x min and x max is the minimum and maximum value of the x coordinate after projection, and y min and max The minimum and maximum values ​​of the projected y coordinate; image_resolution is the resolution of the image, that is, the number of grids. The calculation formula of the area A of the connected domain is: A = n·pixel size , where n is the number of buildings.

4. The method for processing radar scanning point cloud data of power corridor lines according to claim 2 is characterized in that: The calculation formula for the area of ​​each cluster in step f-6 is: In the formula, S represents the area of ​​the polygon; n represents the number of polygon vertices, that is, the number of cluster boundary points; (x i ,y i ) is the coordinate of the ith vertex, (x i+1 ,y i+1 ) are the coordinates of the i+1th vertex.

5. The method for processing radar scanning point cloud data of power corridor lines according to claim 2 is characterized in that: The method for reassigning the Z coordinate of the cluster point cloud obtained in step f-6 is: Find the nearest neighbor of the sampling point in the original point cloud and calculate the average value of its Z coordinate according to the following formula: new And assign it to all points in the cluster, In the formula, z new is the average value of the new z coordinates in the cluster, N represents the number of points in the cluster, and z i is the z coordinate of the ith point in the nearest neighbor.

6. A system based on the method for processing radar scanning point cloud data of the power corridor line according to claim 1, characterized in that: The user first initiates a request to the backend through the frontend page, and the request is received and processed by the high-performance HTTP server interface; The high-performance HTTP server interface interacts with the point cloud database obtained after point cloud data processing, and queries the directory and path information of the required files through the point cloud database; After the point cloud database returns the query result, the high-performance HTTP server interface obtains the file path or actual file content and returns it to the front-end page, or the high-performance HTTP server interface obtains the actual file through the file path and generates a report file; End users download point cloud files or report files through the page.

7. The system of the method for processing radar scanning point cloud data of power corridor lines according to claim 6 is characterized in that: During the file transfer process, multiple levels of monitoring are used, including Level 1 monitoring: detect that both the las point cloud file and the kml file are transferred to the specified folder, and the startup code starts running; Secondary monitoring: After the backend code is finished, las files with different segmentation results are generated in the specified folder, and the "Start potreeconverter" conversion tool is started to convert the point cloud format; Level 3 monitoring: monitor the folder where the conversion is completed. When a point cloud file comes in, the corresponding information of the point cloud file is stored in the database.

Images