Local power transmission line near-electricity early warning method based on laser point cloud data

Abstract

The invention discloses an in-situ power transmission line near-electricity early warning method based on laser point cloud data, which belongs to the field of power detection, and performs down-sampling processing on point cloud data through a voxel grid down-sampling algorithm so as to reduce the number of point clouds, improve the processing efficiency and scan and process the point cloud data of a moving object in real time. And after data are segmented through a DBSCAN algorithm, the nearest distance between the moving object and the power transmission line is calculated by using a Haversine formula and the Pythagorean theorem. The method mainly comprises the following steps: acquiring and processing point cloud data of a power transmission line; acquiring laser radar data; removing noise points; performing down-sampling processing; storing data; real-time point cloud data acquisition: the laser radar scans the moving object below the road in real time to obtain the point cloud data of the moving object. The method is mainly used for near-electricity early warning of the in-situ power transmission line.

Description

Technical Field

[0001] The present invention relates to the field of electric power detection, and in particular to an on-site transmission line near-electricity early warning method based on laser point cloud data. Background Art

[0002] With the development of social economy, the safe and stable operation of the power system as the infrastructure of modern society is of vital importance. In the operation and maintenance of transmission lines, ensuring the safety of the area below the line is of utmost importance, especially in dense areas such as highways, railways, and urban buildings. Moving objects (such as vehicles, mechanical equipment, or pedestrians) may enter the safety range of the transmission line, posing serious safety hazards. The traditional manual inspection method is inefficient, has blind spots and delays, and is difficult to meet the efficient operation requirements of modern power systems.

[0003] LiDAR technology is gradually being used in the safety monitoring of power transmission lines due to its high-precision three-dimensional spatial mapping capabilities. The three-dimensional point cloud data obtained by LiDAR can reflect the conditions of the power transmission line and its surrounding environment in detail. For example, the publication number CN111952883B discloses a power transmission line fault identification system and method based on three-dimensional LiDAR, but it does not involve algorithm improvements for local power transmission line near-electricity warning. Summary of the invention

[0004] The purpose of the present invention is to provide an on-site transmission line near-electricity warning method based on laser point cloud data, which downsamples the point cloud data through a voxel grid downsampling algorithm to reduce the number of point clouds and improve processing efficiency; scans and processes the point cloud data of a moving object in real time, segments the data through the DBSCAN algorithm, and uses the Haversine formula and the Pythagorean theorem to calculate the closest distance between the moving object and the transmission line. When the distance is less than a preset threshold, the system promptly issues a warning signal.

[0005] The present invention is achieved through the following technical solutions:

[0006] A local transmission line near-electricity early warning method based on laser point cloud data comprises the following steps:

[0007] S1. Transmission line point cloud data collection and processing:

[0008] (1) LiDAR data collection: Use LiDAR equipment to scan the power lines above the road and obtain point cloud data;

[0009] (2) Noise point removal: Use the mean filter algorithm to filter the collected point cloud data to remove outliers and invalid data points to ensure the accuracy of the stored data;

[0010] (3) Downsampling processing;

[0011] (4) Data storage: The processed transmission line point cloud data is stored in the edge computing device as the reference data for subsequent calculations. This data includes the precise latitude and longitude coordinates and height information of the transmission line.

[0012] S2. Real-time collection and processing of moving object point cloud data:

[0013] (1) Real-time point cloud data collection: The LiDAR scans the moving objects under the road in real time to obtain their point cloud data;

[0014] (2) Data preprocessing: Similar to the data processing of power transmission lines, noise filtering is performed on the point cloud data collected in real time to ensure data quality; similar to the data processing of power transmission lines, voxel grid downsampling is performed on the point cloud data of moving objects to reduce the number of point clouds and improve processing efficiency;

[0015] S3, data classification and segmentation: Use the DBSCAN algorithm to segment the real-time point cloud data and extract the point cloud data of moving objects;

[0016] S4, coordinate conversion: in point cloud data processing, the rectangular coordinate system of the point is converted into latitude and longitude coordinates and altitude coordinates;

[0017] S5. Distance calculation: including horizontal distance and vertical distance;

[0018] S6. Early warning judgment.

[0019] Furthermore, noise point removal also includes:

[0020] 1) Define the neighborhood:

[0021] For each point P in the point cloud data i =(x i ,y i ,z i ), define a neighborhood N(P i ), the neighborhood contains point P i Other points within a certain range. This range can be determined by distance or a fixed number of nearest neighbors;

[0022] 2) Calculate the distance between the points in the neighborhood and the current point:

[0023] Calculate the value of point Pi and each point P in its neighborhood j =(x j ,y j ,z j ) is:

[0024]

[0025] In the formula, dist(P i ,P j ) represents point P i and point P j The distance between

[0026] 3) Calculate the average distance of points in the neighborhood:

[0027] Calculate point P i The average distance D to all points in its neighborhood avg :

[0028]

[0029] Where n is the neighborhood N(P i )

[0030] 4) Calculate the standard deviation of the distance:

[0031] Calculate point P i The standard deviation of the distances between all points in its neighborhood

[0032]

[0033] In the formula, Representative point P i The degree of dispersion of the distance between it and other points in its neighborhood;

[0034] 5) Noise point judgment:

[0035] According to the average value of the distance D avg and standard deviation Judgment point P i Whether it is a noise point; if the distance between a point and other points in its neighborhood is significantly greater than the average distance plus one or two standard deviations, the point is marked as a noise point;

[0036] like Then point P i is a noise point.

[0037] Where k is a constant, usually 1 or 2, which is used to control the strictness of noise detection.

[0038] Furthermore, the downsampling process also includes: downsampling the point cloud data into voxel grids to reduce the number of point clouds and improve the efficiency of subsequent processing.

[0039] Voxel grid downsampling algorithm:

[0040]

[0041] Where: P voxelis the average point in the voxel, that is, the representative point after downsampling; V is the voxel grid, whose size is determined by voxel_size; |V| is the number of points in the voxel; P i is the point in the original point cloud that is inside voxel V.

[0042] Furthermore, the DBSCAN algorithm includes:

[0043] 1) For each point, find all the points in its neighborhood. If the number of points in the neighborhood is greater than the set minimum number of samples, mark it as a core point;

[0044] 2) Classify the core point and all points in its neighborhood into the same cluster;

[0045] 3) If a point is not a core point and does not belong to any cluster, it is marked as a noise point;

[0046] 4) Repeat the above steps for all unprocessed points until all points have been processed.

[0047] Further, the coordinate conversion includes: converting the rectangular coordinate system (x, y, z) of the point into longitude and latitude coordinates (φ, λ) and altitude coordinates (h);

[0048] 1) Latitude φ calculation:

[0049]

[0050] Where x, y, z are the original coordinates in the point cloud data, and φ is the converted latitude;

[0051] 2) Longitude λ calculation:

[0052]

[0053] In the formula, x, y are the original coordinates in the point cloud data, and λ is the converted longitude;

[0054] 3) Calculation of height h:

[0055] The height h can be obtained directly from the z coordinate in the original rectangular coordinate system:

[0056] h=z

[0057] Where z is the height coordinate in the point cloud data.

[0058] Furthermore, the horizontal distance calculation in the distance calculation is:

[0059]

[0060] Where: d is the horizontal distance between two points; r is the radius of the earth (about 6371km); φ 1 ,λ1 is the latitude and longitude of the moving object; φ 2 ,λ 2 is the latitude and longitude of the transmission line;

[0061] Vertical distance calculation in distance calculation: calculate the difference between the height of the moving object and the height of the transmission line.

[0062] h diff =h 1 -h 2

[0063] Where: h diff is the vertical distance;

[0064] Calculation of total distance: Calculate the total distance between the moving object and the transmission line by the sum of the squares of the horizontal distance and the vertical distance;

[0065]

[0066] Where: D total is the total distance between the moving object and the transmission line; d is the horizontal distance between the two points; h diff is the vertical distance between two points.

[0067] Furthermore, the early warning judgment includes: calculating the total distance D total The preset safety distance threshold D threshold For comparison, if D total Less than or equal to D threshold , then a warning signal is triggered.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] 1. Obtain high-precision three-dimensional point cloud data through lidar, and combine it with advanced algorithms to accurately process the data to ensure the accuracy of the monitoring results. It is suitable for accurate monitoring in complex environments.

[0070] 2. It realizes real-time monitoring and distance calculation of moving objects under the transmission lines, can quickly identify potential dangers, and issue early warnings in time to effectively prevent accidents.

[0071] 3. It has all-weather, automated monitoring capabilities, can operate stably under various environmental conditions, reduce human intervention, achieve continuous safety monitoring of transmission lines, and ensure system reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 is a flow chart of the present invention;

[0073] Figure 2It is a structural schematic diagram of the local type power transmission line near-electricity warning device of the laser point cloud data of the present invention;

[0074] Figure 3 is a circuit diagram of the boost module of the present invention.

[0075] In the picture: 1. LiDAR; 2. 4G communication module; 3. Lithium battery; 4. RTK module; 5. RTK antenna. DETAILED DESCRIPTION

[0076] The present invention will be further described below in conjunction with the accompanying drawings.

[0077] Example 1

[0078] A local transmission line near-electricity early warning method based on laser point cloud data comprises the following steps:

[0079] S1. Transmission line point cloud data collection and processing:

[0080] (1) LiDAR data collection: Use LiDAR equipment to scan the power lines above the road and obtain point cloud data;

[0081] (2) Noise point removal: Use the mean filter algorithm to filter the collected point cloud data to remove outliers and invalid data points to ensure the accuracy of the stored data;

[0082] (3) Downsampling processing;

[0083] (4) Data storage: The processed transmission line point cloud data is stored in the edge computing device as the reference data for subsequent calculations. This data includes the precise latitude and longitude coordinates and height information of the transmission line.

[0084] S2. Real-time collection and processing of moving object point cloud data:

[0085] (1) Real-time point cloud data acquisition: The LiDAR scans moving objects (such as pedestrians and vehicles) under the road in real time to obtain their point cloud data;

[0086] (2) Data preprocessing: Similar to the data processing of power transmission lines, noise filtering is performed on the point cloud data collected in real time to ensure data quality; similar to the data processing of power transmission lines, voxel grid downsampling is performed on the point cloud data of moving objects to reduce the number of point clouds and improve processing efficiency;

[0087] S3, data classification and segmentation: Use the DBSCAN algorithm to segment the real-time point cloud data and extract the point cloud data of moving objects;

[0088] S4, coordinate conversion: in point cloud data processing, the rectangular coordinate system of the point is converted into latitude and longitude coordinates and altitude coordinates;

[0089] S5. Distance calculation: including horizontal distance (distance along the earth's surface) and vertical distance (height difference);

[0090] S6. Early warning judgment.

[0091] Example 2

[0092] A local transmission line near-electricity early warning method based on laser point cloud data, wherein noise point removal also includes:

[0093] 1) Define the neighborhood:

[0094] For each point P in the point cloud data i =(x i ,y i ,z i ), define a neighborhood N(P i ), the neighborhood contains point P i Other points within a certain range. This range can be determined by distance or a fixed number of nearest neighbors;

[0095] 2) Calculate the distance between the points in the neighborhood and the current point:

[0096] Calculate the value of point Pi and each point P in its neighborhood j =(x j ,y j ,z j ) is:

[0097]

[0098] In the formula, dist(P i ,P j ) represents point P i and point P j The distance between

[0099] 3) Calculate the average distance of points in the neighborhood:

[0100] Calculate point P i The average distance D to all points in its neighborhood avg :

[0101]

[0102] Where n is the neighborhood N(P i )

[0103] 4) Calculate the standard deviation of the distance:

[0104] Calculate point P i The standard deviation of the distances between all points in its neighborhood

[0105]

[0106] In the formula, Representative point P i The degree of dispersion of the distance between it and other points in its neighborhood;

[0107] 5) Noise point judgment:

[0108] According to the average value of the distance D avg and standard deviation Judgment point P i Whether it is a noise point; if the distance between a point and other points in its neighborhood is significantly greater than the average distance plus one or two standard deviations, the point is marked as a noise point;

[0109] like Then point P i is a noise point.

[0110] Where k is a constant, usually 1 or 2, which is used to control the strictness of noise detection;

[0111] The downsampling process also includes: downsampling the point cloud data into voxel grids to reduce the number of point clouds and improve the efficiency of subsequent processing. The voxel grid is a grid structure that divides the three-dimensional space into a series of cubes (voxels). Each voxel represents a fixed-size cubic area.

[0112] Voxel grid downsampling algorithm:

[0113]

[0114] Where: P voxel is the average point in the voxel, that is, the representative point after downsampling; V is the voxel grid, whose size is determined by voxel_size; |V| is the number of points in the voxel; P i is the point within the voxel V in the original point cloud; voxel_size is the size of the voxel grid, which determines the accuracy of downsampling. A larger voxel_size will result in a coarser downsampling (reducing more points), while a smaller voxel_size will retain more details;

[0115] The DBSCAN algorithm includes:

[0116] 1) For each point, find all the points in its neighborhood. If the number of points in the neighborhood is greater than the set minimum number of samples (min_samples), mark it as a core point;

[0117] 2) Classify the core point and all points in its neighborhood into the same cluster;

[0118] 3) If a point is not a core point and does not belong to any cluster, it is marked as a noise point;

[0119] 4) Repeat the above steps for all unprocessed points until all points have been processed;

[0120] Coordinate transformation includes: converting the rectangular coordinate system (x, y, z) of the point into longitude and latitude coordinates (φ, λ) and altitude coordinates (h);

[0121] 1) Latitude φ calculation:

[0122]

[0123] Where x, y, z are the original coordinates in the point cloud data, and φ is the converted latitude;

[0124] 2) Longitude λ calculation:

[0125]

[0126] In the formula, x, y are the original coordinates in the point cloud data, and λ is the converted longitude;

[0127] 3) Calculation of height h:

[0128] The height h can be obtained directly from the z coordinate in the original rectangular coordinate system:

[0129] h=z

[0130] Where z is the height coordinate in the point cloud data;

[0131] Horizontal distance calculation in distance calculation:

[0132] The Haversine formula is used to calculate the horizontal distance between the highest point of the moving object and the transmission line. The Haversine formula takes into account the curvature of the earth and is a formula for calculating the great circle distance between two points.

[0133]

[0134] Where: d is the horizontal distance between two points; r is the radius of the earth (about 6371km); φ 1 ,λ 1 is the latitude and longitude of the moving object; φ 2 ,λ 2 is the latitude and longitude of the transmission line;

[0135] Vertical distance calculation in distance calculation: calculate the difference between the height of the moving object and the height of the transmission line.

[0136] h diff =h1 -h 2

[0137] Where: h diff is the vertical distance;

[0138] Calculation of total distance: Calculate the total distance (spatial distance) between the moving object and the transmission line by the sum of the squares of the horizontal distance and the vertical distance;

[0139]

[0140] Where: D total is the total distance between the moving object and the transmission line; d is the horizontal distance between the two points; h diff is the vertical distance between two points;

[0141] The early warning judgment includes: the calculated total distance D total The preset safety distance threshold D threshold For comparison, if D total Less than or equal to D threshold , then a warning signal is triggered, and the rest is the same as in Example 1.

[0142] like Figure 2 As shown, the early warning method is implemented based on the local transmission line near-electricity early warning device of laser point cloud data, including a shell, a control unit, a buzzer, a temperature and humidity sensor and a lithium battery 3 are arranged in the shell, a laser radar 1 is arranged above the shell, and an RTK antenna 5 is arranged below the shell. The control unit is respectively connected to the 4G communication module 2 and the RTK module 4, and the RTK module 4 is connected to the RTK antenna 5. The control unit is also connected to the laser radar 1, the lithium battery 3 and the temperature and humidity sensor, and the temperature and humidity sensor is used to detect the temperature and humidity of the environment. The local transmission line near-electricity early warning device of laser point cloud data is respectively connected to the USB and the lithium battery 3 through the power battery management module. The power battery management chip adopts IP2312U_VSET, and the VBAT pin of IP2312U_VSET is connected to the lithium battery 3. The battery power is used as the power supply of the main control unit, and the control unit leads to a 3.3V power supply as the power supply of the RTK module. The lithium battery 3 is connected to the buzzer through the boost module.

[0143] like Figure 3As shown, the boost module includes a boost chip U1, the SW pin of the boost chip U1 is connected to a diode D3, the FB pin of the boost chip U1 is respectively connected to a voltage divider resistor R9 and an adjustment resistor R8, the diode D3 is connected to filter capacitors C2 and C3, the VCC pin of the boost chip U1 is connected to filter capacitors C4, C5, and C6, and an inductor L3 is connected between the VCC pin and the SW pin of the boost chip U1. The boost chip U1 uses sx1308, and R8 uses a 191K resistor when the boost value is 12V, and R18 uses a 73.2K resistor when the boost value is 5V.

[0144] like Figure 1 As shown, the device is connected to the 4G network, then connected to the remote cloud server, and uses the mobile RTK account to connect to the Zhicun Space CORS service to obtain centimeter-level high-precision positioning, and then obtains the temperature and humidity from the temperature and humidity sensor. The above process is the positioning obtained by the device itself and the various data of auxiliary positioning. The control unit controls the laser radar 1 to detect the object data within the space range, and transmits these data and the device split positioning information data, wire point cloud data and temperature and humidity sensor data to the processor of the control unit. The processor performs edge computing to obtain the distance from the wire to each object; the threshold value of the acquired distance data is judged, and if it exceeds the threshold, an alarm is triggered, otherwise the data is transmitted to the cloud platform.

[0145] Equipment working steps:

[0146] (1) Import the conductor point cloud data into this system to generate the octocmap of the transmission line;

[0147] (2) After the device is installed and powered on, the device's geographic location and the spatial information detected by the radar inside the device are transmitted to the high-performance processor;

[0148] (3) The high-performance processor in the device performs edge computing, converting data such as the device location, wire location, and the location of other objects within the radar monitoring range into a spatial model in the ROS system, and quickly performs collision detection calculations through Moveit's FCL (FlexibleCollision Library);

[0149] (4) Determine whether to issue an early warning based on the calculation results and upload the information to the cloud platform.

Claims

1. A local transmission line near-electricity early warning method based on laser point cloud data, characterized in that: The following steps are included: S1. Transmission line point cloud data collection and processing: (1) LiDAR data collection; (2) Noise point removal: Use the mean filter algorithm to filter the collected point cloud data to remove outliers and invalid data points to ensure the accuracy of the stored data; (3) Downsampling processing; (4) Data storage; S2. Real-time collection and processing of moving object point cloud data: (1) Real-time point cloud data collection: The LiDAR scans the moving objects under the road in real time to obtain their point cloud data; (2) Data preprocessing: Noise filtering is performed on point cloud data collected in real time to ensure data quality; voxel grid downsampling is performed on point cloud data of moving objects to reduce the number of point clouds and improve processing efficiency; S3, data classification and segmentation: Use the DBSCAN algorithm to segment the real-time point cloud data and extract the point cloud data of moving objects; S4, coordinate conversion: in point cloud data processing, the rectangular coordinate system of the point is converted into latitude and longitude coordinates and altitude coordinates; S5. Distance calculation: including horizontal distance and vertical distance; S6. Early warning judgment.

2. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 1 is characterized by: Noise point removal also includes: 1) Define the neighborhood: For each point P in the point cloud data i =(x i ,y i ,z i ), define a neighborhood N(P i ), the neighborhood contains point P i Other points within a certain range, determined by distance or a fixed number of nearest neighbors; 2) Calculate the distance between the points in the neighborhood and the current point: Calculate the value of point Pi and each point P in its neighborhood j =(x j ,y j ,z j ), In the formula, dist(P i ,P j ) represents point P i and point P j The distance between 3) Calculate the average distance of points in the neighborhood: Calculate point P i The average distance D to all points in its neighborhood avg , Where n is the neighborhood N(P i ) The number of points within ; 4) Calculate the standard deviation of the distance: Calculate point P i The standard deviation of the distances between all points in its neighborhood In the formula, Representative point P i The degree of dispersion of the distance between it and other points in its neighborhood; 5) Noise point judgment: According to the average value of the distance D avg and standard deviation Judgment point P i Whether it is a noise point; if the distance between a point and other points in its neighborhood is significantly greater than the average distance plus one or two standard deviations, the point is marked as a noise point; like Then point P i is a noise point. Where k is a constant, usually 1 or 2, which is used to control the strictness of noise detection.

3. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 1 is characterized by: Downsampling processing also includes: downsampling the point cloud data by voxel grid to reduce the number of point clouds and improve the efficiency of subsequent processing. Voxel grid downsampling algorithm: Where: P voxel is the average point in the voxel, that is, the representative point after downsampling; V is the voxel grid, whose size is determined by voxel_size; |V| is the number of points in the voxel; P i is the point in the original point cloud that is inside voxel V.

4. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 1 is characterized by: The DBSCAN algorithm includes: 1) For each point, find all the points in its neighborhood. If the number of points in the neighborhood is greater than the set minimum number of samples, mark it as a core point; 2) Classify the core point and all points in its neighborhood into the same cluster; 3) If a point is not a core point and does not belong to any cluster, it is marked as a noise point; 4) Repeat the above steps for all unprocessed points until all points have been processed.

5. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 1 is characterized by: Coordinate transformation includes: converting the rectangular coordinate system (x, y, z) of the point into longitude and latitude coordinates (φ, λ) and altitude coordinates (h); 1) Latitude φ calculation: Where x, y, z are the original coordinates in the point cloud data, and φ is the converted latitude; 2) Longitude λ calculation: In the formula, x, y are the original coordinates in the point cloud data, and λ is the converted longitude; 3) Calculation of height h: The height h is obtained directly from the z coordinate in the original rectangular coordinate system: h=z Where z is the height coordinate in the point cloud data.

6. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 1 is characterized by: Horizontal distance calculation in distance calculation: Where: d is the horizontal distance between two points; r is the radius of the earth (about 6371km); φ1, λ1 are the latitude and longitude of the moving object; φ2, λ2 are the latitude and longitude of the transmission line; Vertical distance calculation in distance calculation: calculate the difference between the height of the moving object and the height of the transmission line. h diff =h1-h2 Where: h diff is the vertical distance; Calculation of total distance: The total distance between the moving object and the transmission line is calculated by the sum of the squares of the horizontal distance and the vertical distance. Where: D total is the total distance between the moving object and the transmission line; d is the horizontal distance between the two points; h diff is the vertical distance between two points.

7. The local power transmission line near-electricity early warning method based on laser point cloud data according to claim 6 is characterized by: The early warning judgment includes: the calculated total distance D total The preset safety distance threshold D threshold For comparison, if D total Less than or equal to D threshold , then a warning signal is triggered.

Images