Power transmission line tree obstacle detection method based on thunder-vision fusion

Abstract

The invention belongs to the field of power transmission line tree obstacle detection, and particularly relates to a power transmission line tree obstacle detection method based on thunder-vision fusion, and the method comprises a data collection module which comprehensively collects the original data of the surrounding environment of a power transmission line through a laser radar, a camera and POS equipment. The advantages are obvious. Laser radar and camera data are fused, data registration and feature extraction technologies are applied, tree obstacles are accurately identified and positioned, the danger degree is accurately judged, and the detection precision is far higher than that of a single sensor. Quasi-real-time data transmission is achieved through the point cloud technology, and hidden dangers can be found in time in cooperation with real-time modeling and rapid distance measurement. The intelligent device integrates multiple sensors, automatically analyzes tree barrier risks, reduces manual intervention and reduces cost. In addition, detailed detection reports can be generated, support is provided for operation and maintenance, multi-element perception fusion enhances environmental adaptability, and stable operation of the power transmission line is guaranteed.

Description

Technical Field

[0001] The invention belongs to the field of tree obstacle detection for power transmission lines, and in particular to a tree obstacle detection method for power transmission lines based on radar and visual fusion. Background Art

[0002] In the modern power transmission system, transmission lines are the key channels for power transmission, and their safe and stable operation is of vital importance. However, transmission lines are usually widely distributed, passing through various complex terrains and environments, and tree obstacles caused by tree growth have become one of the common hidden dangers threatening the safety of transmission lines.

[0003] Traditional means of detecting tree obstacles on power transmission lines have many limitations. Manual inspections are not only inefficient, but also greatly affected by terrain, weather, and the experience and subjective factors of inspectors. It is difficult to fully and timely discover hidden dangers of tree obstacles, especially in mountainous areas, jungles and other areas with complex terrain. The difficulty and risk of manual inspections are greatly increased. Some early detection technologies based on single sensors, such as relying solely on cameras for image acquisition and identification of tree obstacles, are easily disturbed by changes in light and weather. In severe weather conditions (such as heavy rain, fog, night, etc.), the image quality is seriously degraded, resulting in a significant reduction in detection accuracy or even failure to work properly. When only LiDAR is used for detection, although relatively accurate three-dimensional spatial information can be obtained, LiDAR data lacks texture and color information, and it is difficult to accurately identify tree species and judge tree obstacle risks based on texture features, making it difficult to meet the needs of comprehensive and accurate detection of tree obstacles.

[0004] With the continuous growth of electricity demand and the continuous expansion of the power grid, higher requirements are placed on the reliability and stability of the safe operation of transmission lines. The shortcomings of traditional detection methods are becoming more and more prominent, and an efficient, accurate and adaptable tree obstacle detection technology is urgently needed.

[0005] To this end, the present invention provides a transmission line tree obstacle detection method based on radar and vision fusion. Summary of the invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve the technical problem is: the tree obstacle detection method for power transmission lines based on radar and visual fusion of the present invention is characterized by comprising:

[0008] The data acquisition module uses laser radar, cameras and POS devices to comprehensively collect the original data of the surrounding environment of the transmission line, providing basic data support for subsequent links;

[0009] The data preprocessing module denoises and filters lidar data, and grayscales, normalizes, and denoises visual images to improve data quality;

[0010] The feature extraction and matching module extracts key features from lidar and visual image data and establishes corresponding relationships to mutually verify the two types of data, providing key information for accurate detection of tree obstacles;

[0011] The tree obstacle detection and analysis module calculates the distance between trees and transmission lines, divides the tree obstacle levels based on factors such as tree species, size, and distance, analyzes historical data to predict the growth trend of trees, and provides a direct basis for preventive measures;

[0012] The result output and warning module generates a report on the detection results, sets a warning threshold, and issues a timely warning when a dangerous situation occurs;

[0013] The point cloud acquisition, modeling, compression, and transmission module quickly acquires point cloud data, models and compresses it using fast incremental triangular meshing technology, realizes quasi-real-time transmission, improves data processing efficiency, and provides strong data support for subsequent work;

[0014] The multi-sensor information fusion module uses 2D-2D, 3D-3D, 3D-2D data registration algorithms and multi-sensor clock synchronization technology to highly accurately fuse lidar and visible light camera data, providing precise measurement data for tree obstacle ranging and analysis;

[0015] The lidar-vision intelligent device system module integrates a lidar, a visible light camera, and a POS sensor, and uses real-time data modeling and fast ranging technology to realize intelligent and real-time detection and analysis of tree obstacles on transmission lines, improving detection accuracy and efficiency.

[0016] Furthermore, the data acquisition module mainly consists of the following parts:

[0017] The lidar acquires three-dimensional laser point clouds, accurately outlining the terrain undulations, tree heights and positions around the transmission line, as well as the spatial layout of the poles and conductors, providing an accurate geometric basis for subsequent distance measurement and spatial analysis;

[0018] The camera acquires visible light video images, clearly presenting the tree species, bark texture, leaf morphology and color, helping to identify the growth characteristics of different tree species and assisting in judging the tree obstacle risk;

[0019] The POS device obtains the device attitude information, records the real-time position and angle of the device when collecting data, ensures the spatial consistency of the laser point cloud and the video image, and enables subsequent data fusion and analysis to be carried out in a unified coordinate system.

[0020] Furthermore, the data preprocessing module mainly consists of the following parts:

[0021] Denoise and filter lidar data to remove outliers caused by measurement errors and electromagnetic interference, reduce data redundancy, improve the accuracy and stability of the data, and make subsequent feature extraction more reliable;

[0022] Grayscale, normalize, and denoise visual images. Grayscaling facilitates the processing of image feature extraction algorithms; normalization makes images under different lighting conditions have a unified standard and enhances image comparability; denoising removes noise points in the image, improves image clarity, and provides high-quality image data for subsequent image recognition and analysis.

[0023] Furthermore, the feature extraction and matching module mainly consists of the following parts:

[0024] Extract features from lidar data to obtain accurate geometric features of trees and transmission lines, such as the volume of trees and the curvature of transmission lines, providing a quantitative basis for distance calculation and tree obstacle analysis;

[0025] Extract features from visual images to capture the texture and color visual features of trees, providing intuitive visual information for tree species identification and tree obstacle classification, and supplementing the deficiencies of lidar data in texture and color information;

[0026] Feature matching to establish a spatial correspondence between lidar data and visual images, enabling the two types of data to verify and supplement each other, and improving the accuracy and reliability of tree obstacle detection.

[0027] Furthermore, the tree obstacle detection and analysis module mainly consists of the following parts:

[0028] Calculate the distance between trees and transmission lines, accurately measure the actual distance between the two, provide a quantitative standard for judging the degree of danger of tree obstacles, and determine whether corresponding measures need to be taken;

[0029] Identify and classify tree obstacles. According to factors such as tree species, size, and distance, classify tree obstacles into different levels to facilitate maintenance personnel to formulate targeted treatment plans;

[0030] Analyze the growth trend of trees. Through the analysis of historical data, predict the future growth direction and speed of trees, early warning of potential tree obstacle risks, and providing a decision-making basis for early prevention.

[0031] Furthermore, the result output and warning module mainly consists of the following parts:

[0032] Generate a tree obstacle detection report, detailedly record various data and analysis results of tree obstacle detection, providing comprehensive data support and historical reference for operation and maintenance management;

[0033] Set the warning threshold and issue warnings and alarms. When the distance or growth trend of tree obstacles reaches the danger threshold, send warning and alarm signals to the operation and maintenance personnel in a timely manner for a quick response to avoid accidents.

[0034] Furthermore, the point cloud acquisition, modeling, compression, and transmission module mainly consists of the following parts:

[0035] Point cloud acquisition, quickly and comprehensively obtain the three-dimensional point cloud data of the surrounding environment of the transmission line, providing a rich data source for subsequent modeling and analysis;

[0036] Point cloud modeling, using fast incremental triangulation technology to efficiently construct a high-precision three-dimensional model, intuitively showing the spatial relationship between the transmission line and the surrounding environment;

[0037] Point cloud compression, adopting the V-PCC dynamic point cloud compression method to reduce the data volume without losing key information, facilitating data storage and transmission;

[0038] Point cloud transmission, realizing quasi-real-time point cloud data transmission, enabling operation and maintenance personnel to obtain the latest information on the surrounding environment of the transmission line in a timely manner, providing support for real-time monitoring and analysis.

[0039] Furthermore, the multi-sensor information fusion module mainly consists of the following parts:

[0040] Using data registration algorithms and clock synchronization technologies, achieve high-precision registration of lidar and visible light camera data, eliminate data deviations caused by equipment time asynchronization and spatial position differences, and improve the accuracy of data fusion;

[0041] Data fusion, complement the advantages of the two types of data to obtain richer and more accurate point cloud measurement data, providing more reliable theoretical support for on-site rapid tree obstacle distance measurement and analysis. The main components of the radar-vision intelligent device system module include the following parts:

[0042] Integrate multiple sensors to achieve all-round perception of the surrounding environment of the transmission line, integrate the data of different sensors, and provide a comprehensive data basis for subsequent analysis;

[0043] Real-time data modeling and rapid distance measurement technology, real-time construct the model of the transmission line and the surrounding environment, and quickly calculate the distance between the tree and the transmission line to achieve real-time monitoring and analysis of tree obstacles;

[0044] Carry out intelligent distance measurement analysis of the safety distance, automatically judge whether the tree obstacle threatens the safety of the transmission line, improve the intelligence, real-time performance, and accuracy of tree obstacle detection, and reduce the cost and risk of manual detection.

[0045] A method for detecting tree obstacles on transmission lines based on radar-vision fusion, the detection method includes the following steps:

[0046] S1. Data acquisition: Use lidar to scan the transmission line and its surroundings to obtain 3D lidar point clouds and determine spatial positions; collect visible light video images through cameras to capture tree textures and colors; use POS devices to obtain real-time position and angle information of the devices to ensure data spatial consistency;

[0047] S2. Data preprocessing: Denoise and filter the lidar data to improve data accuracy; successively grayscale, normalize, and denoise the visual images to enhance image quality for subsequent processing;

[0048] S3. Point cloud processing and transmission: Rapidly collect 3D point cloud data around the transmission line, model using fast incremental triangular meshing technology, compress the data using the V-PCC method, and achieve quasi-real-time transmission to provide support for subsequent analysis;

[0049] S4. Feature extraction and matching: Extract geometric features of trees and transmission lines from lidar data, extract texture and color visual features from visual images, and then perform feature matching to improve the accuracy of tree obstacle detection;

[0050] S5. Multi-sensory information fusion: Use 2D-2D, 3D-3D, 3D-2D data registration algorithms and multi-sensor clock synchronization technology to accurately register and fuse lidar and visible light camera data, providing theoretical support for tree obstacle ranging and analysis;

[0051] S6. Tree obstacle detection and analysis: Calculate the distance between the tree and the transmission line, classify the tree obstacle levels according to the tree species, size, and distance, analyze historical data to predict the tree growth trend, and judge the danger level of the tree obstacle;

[0052] S7. Result output and warning: Organize the detection data and analysis results to generate reports to provide data support for operation and maintenance; set warning thresholds, and when the tree obstacle reaches the danger threshold, promptly warn the operation and maintenance personnel to avoid accidents.

[0053] The beneficial effects of the present invention are as follows:

[0054] 1. The method for detecting tree obstacles on transmission lines based on the fusion of lidar and vision according to the present invention, through multi-source data fusion, the accurate spatial position information obtained by lidar and the rich texture and color information captured by the camera complement each other. Combining 2D-2D, 3D-3D, 3D-2D data registration algorithms and feature extraction and matching technologies, it realizes the precise identification and positioning of tree obstacles, can accurately calculate the distance between the tree and the transmission line, accurately judge the danger level of the tree obstacle, and the accuracy of classification and analysis far exceeds that of single-sensor detection.

[0055] 2. The tree fault detection method for transmission lines based on thunder and vision fusion according to the present invention uses fast incremental triangular meshing technology through the point cloud acquisition, modeling, compression, and transmission module to achieve quasi-real-time point cloud data transmission. In cooperation with data preprocessing, real-time data modeling, and fast distance measurement technology, the time from data acquisition to the output of the analysis result is greatly shortened, potential tree fault hazards can be discovered in a timely manner, and it is convenient for quick response and processing.

[0056] 3. The tree fault detection method for transmission lines based on thunder and vision fusion according to the present invention integrates a variety of sensors through the thunder and vision intelligent device system module, uses intelligent algorithms to carry out intelligent distance measurement analysis of safety distances, automatically judges tree fault risks, reduces manual intervention, lowers the costs and risks of manual detection, improves the intelligence level of tree fault detection, and meets the intelligent requirements of modern transmission line operation and maintenance.

[0057] 4. The tree fault detection method for transmission lines based on thunder and vision fusion according to the present invention generates a detailed tree fault detection report, covering information such as the location, type, danger level, and growth trend of tree faults, provides comprehensive data support for operation and maintenance management, facilitates the formulation of scientific and reasonable preventive measures and maintenance plans, and ensures the long-term stable operation of transmission lines.

[0058] 5. The tree fault detection method for transmission lines based on thunder and vision fusion according to the present invention can effectively integrate data from different types of sensors through the multi-sensor information fusion module, reduce the influence of environmental factors on a single sensor, can work stably in complex environments, and has good adaptability to the detection of tree faults on transmission lines under different terrains and climatic conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The present invention will be further described below with reference to the accompanying drawings.

[0060] Figure 1 is the flowchart of the method in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] In order to make the technical means, creative features, achieved purposes, and functions of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0062] As Figure 1 shown, the tree fault detection method for transmission lines based on thunder and vision fusion in the embodiment of the present invention is characterized by including:

[0063] A data acquisition module, through lidar, cameras, and POS devices, comprehensively acquires the original data of the surrounding environment of the transmission line, providing basic data support for subsequent links;

[0064] The data preprocessing module denoises and filters lidar data, and grayscales, normalizes, and denoises visual images to improve data quality and create favorable conditions for feature extraction;

[0065] The feature extraction and matching module extracts key features from lidar and visual image data and establishes corresponding relationships to mutually verify the two types of data and provide key information for accurate detection of tree obstacles;

[0066] The tree obstacle detection and analysis module calculates the distance between trees and transmission lines, divides the tree obstacle levels based on factors such as tree species, size, and distance, analyzes historical data to predict the growth trend of trees, and provides a direct basis for preventive measures;

[0067] The result output and warning module generates a report on the detection results, sets a warning threshold, and issues a timely warning when a dangerous situation occurs to assist operation and maintenance personnel in quickly responding and ensuring the safety of transmission lines;

[0068] The point cloud acquisition, modeling, compression, and transmission module quickly acquires point cloud data, uses fast incremental triangular meshing technology for modeling and compression, realizes quasi-real-time transmission, improves data processing efficiency, and provides strong data support for subsequent work;

[0069] The multi-sensor information fusion module uses 2D-2D, 3D-3D, and 3D-2D data registration algorithms and multi-sensor clock synchronization technology to highly accurately fuse lidar and visible light camera data and provide precise measurement data for tree obstacle ranging and analysis;

[0070] The lidar-vision intelligent device system module integrates a lidar, a visible light camera, and a POS sensor, and uses real-time data modeling and fast ranging technology to realize intelligent and real-time detection and analysis of tree obstacles on transmission lines, improving detection accuracy and efficiency.

[0071] The data acquisition module mainly consists of the following parts:

[0072] The lidar acquires three-dimensional laser point clouds, accurately outlining the terrain undulations, tree heights and positions around the transmission line, as well as the spatial layout of the poles and conductors, providing a precise geometric basis for subsequent distance measurement and spatial analysis;

[0073] The camera acquires visible light video images, clearly presenting the tree species, bark texture, leaf morphology and color of the trees, helping to identify the growth characteristics of different tree species and assisting in judging the tree obstacle risk;

[0074] The POS device obtains the device attitude information, records the real-time position and angle of the device during data acquisition, ensures the spatial consistency of the laser point cloud and video images, and enables subsequent data fusion and analysis to be carried out in a unified coordinate system.

[0075] Through lidar, cameras, and POS devices, the original data of the surrounding environment of the transmission line is comprehensively collected, providing basic data support for subsequent links. Among them, lidar collects three-dimensional laser point clouds, accurately outlining the terrain undulation, tree height and position around the transmission line, as well as the spatial layout of poles and conductors, providing an accurate geometric basis for subsequent distance measurement and spatial analysis; cameras collect visible light video images, clearly presenting the types of trees, bark textures, leaf shapes and colors, helping to identify the growth characteristics of different tree species and assisting in judging the risk of tree obstacles; the POS device obtains device attitude information, records the real-time position and angle of the device when collecting data, ensures the spatial consistency of the laser point cloud and video images, and enables subsequent data fusion and analysis to be carried out in a unified coordinate system.

[0076] The main components of the data preprocessing module include the following parts:

[0077] Denoise and filter lidar data, remove abnormal points caused by measurement errors and electromagnetic interference, reduce data redundancy, improve the accuracy and stability of the data, and make subsequent feature extraction more reliable;

[0078] Grayscale, normalize, and denoise visual images. Grayscaling facilitates the processing of image feature extraction algorithms; normalization makes images under different lighting conditions have a unified standard, enhancing image comparability; denoising removes noise points in the image, improves image clarity, and provides high-quality image data for subsequent image recognition and analysis.

[0079] Denoise and filter lidar data, and perform grayscaling, normalization, and denoising on visual images to improve data quality and create favorable conditions for feature extraction. Denoise and filter lidar data, remove abnormal points caused by measurement errors and electromagnetic interference, reduce data redundancy, improve the accuracy and stability of the data, and make subsequent feature extraction more reliable; grayscale, normalize, and denoise visual images. Grayscaling facilitates the processing of image feature extraction algorithms; normalization makes images under different lighting conditions have a unified standard, enhancing image comparability; denoising removes noise points in the image, improves image clarity, and provides high-quality image data for subsequent image recognition and analysis.

[0080] The main components of the feature extraction and matching module include the following parts:

[0081] Extract features from lidar data to obtain accurate geometric features of trees and transmission lines, such as the volume of trees and the curvature of transmission lines, providing a quantitative basis for distance calculation and tree obstacle analysis;

[0082] Extract features from visual images to capture the texture and color visual features of trees, providing intuitive visual information for tree species identification and tree obstacle classification, and supplementing the deficiencies of lidar data in texture and color information;

[0083] Feature matching is performed to establish a spatial correspondence between lidar data and visual images, enabling the two types of data to verify and complement each other, thereby improving the accuracy and reliability of tree obstacle detection.

[0084] Extract key features from lidar and visual image data and establish corresponding relationships, enabling the two types of data to verify each other, providing key information for accurate detection of tree obstacles. Extract features from lidar data to obtain precise geometric features of trees and transmission lines, such as the volume of trees and the curvature of transmission lines, providing a quantitative basis for distance calculation and tree obstacle analysis; extract features from visual images to capture the texture and color visual features of trees, providing intuitive visual information for tree species identification and tree obstacle classification, supplementing the deficiencies of lidar data in texture and color information; perform feature matching to establish a spatial correspondence between lidar data and visual images, enabling the two types of data to verify and complement each other, improving the accuracy and reliability of tree obstacle detection.

[0085] The tree obstacle detection and analysis module mainly consists of the following parts:

[0086] Calculate the distance between the tree and the transmission line, accurately measure the actual distance between the two, provide a quantitative standard for judging the degree of danger of the tree obstacle, and determine whether corresponding measures need to be taken;

[0087] Identify and classify tree obstacles. According to factors such as tree species, size, and distance, classify tree obstacles into different levels to facilitate maintenance personnel to formulate targeted treatment plans;

[0088] Analyze the growth trend of trees. Through the analysis of historical data, predict the future growth direction and speed of trees, early warning of potential tree obstacle risks, and providing a decision-making basis for early prevention.

[0089] Calculate the distance between the tree and the transmission line, classify the tree obstacle levels based on factors such as tree species, size, and distance, analyze historical data to predict the growth trend of trees, and provide a direct basis for preventive measures. Calculate the distance between the tree and the transmission line, accurately measure the actual distance between the two, provide a quantitative standard for judging the degree of danger of the tree obstacle, and determine whether corresponding measures need to be taken; identify and classify tree obstacles. According to factors such as tree species, size, and distance, classify tree obstacles into different levels to facilitate maintenance personnel to formulate targeted treatment plans; analyze the growth trend of trees. Through the analysis of historical data, predict the future growth direction and speed of trees, early warning of potential tree obstacle risks, and providing a decision-making basis for early prevention.

[0090] The result output and warning module mainly consists of the following parts:

[0091] Generate a tree obstacle detection report, which details all the data and analysis results of the tree obstacle detection, providing comprehensive data support and historical reference for operation and maintenance management;

[0092] Set warning thresholds and give warnings and alarms. When the distance or growth trend of tree obstacles reaches the danger threshold, send warning and alarm signals to the operation and maintenance personnel in a timely manner for a quick response to avoid accidents.

[0093] Generate reports on the detection results, set warning thresholds, and give warnings in a timely manner when dangerous situations occur to assist the operation and maintenance personnel in making a quick response and ensuring the safety of transmission lines. Generate tree obstacle detection reports, record in detail various data and analysis results of tree obstacle detection, provide comprehensive data support and historical references for operation and maintenance management; set warning thresholds and give warnings and alarms. When the distance or growth trend of tree obstacles reaches the danger threshold, send warning and alarm signals to the operation and maintenance personnel in a timely manner for a quick response to avoid accidents.

[0094] The point cloud acquisition, modeling, compression, and transmission module mainly consists of the following parts:

[0095] Point cloud acquisition, quickly and comprehensively obtain the three-dimensional point cloud data of the surrounding environment of the transmission line, providing a rich data source for subsequent modeling and analysis;

[0096] Point cloud modeling, using technologies such as fast incremental triangular meshing to efficiently construct a high-precision three-dimensional model and visually display the spatial relationship between the transmission line and the surrounding environment;

[0097] Point cloud compression, adopting the V-PCC dynamic point cloud compression method to reduce the data volume without losing key information, facilitating data storage and transmission;

[0098] Point cloud transmission, realizing quasi-real-time point cloud data transmission, enabling the operation and maintenance personnel to obtain the latest information on the surrounding environment of the transmission line in a timely manner and providing support for real-time monitoring and analysis.

[0099] Quickly collect point cloud data, use technologies such as fast incremental triangular meshing for modeling and compression, realize quasi-real-time transmission, improve data processing efficiency, and provide strong data support for subsequent work. Point cloud acquisition, quickly and comprehensively obtain the three-dimensional point cloud data of the surrounding environment of the transmission line, providing a rich data source for subsequent modeling and analysis; point cloud modeling, using technologies such as fast incremental triangular meshing to efficiently construct a high-precision three-dimensional model and visually display the spatial relationship between the transmission line and the surrounding environment; point cloud compression, adopting the V-PCC dynamic point cloud compression method to reduce the data volume without losing key information, facilitating data storage and transmission; point cloud transmission, realizing quasi-real-time point cloud data transmission, enabling the operation and maintenance personnel to obtain the latest information on the surrounding environment of the transmission line in a timely manner and providing support for real-time monitoring and analysis.

[0100] The multi-sensor information fusion module mainly consists of the following parts:

[0101] Using data registration algorithms and clock synchronization technologies, high-precision registration of lidar and visible light camera data is achieved, data deviation caused by device time asynchronization and spatial position differences is eliminated, and the accuracy of data fusion is improved;

[0102] Data fusion combines the advantages of the two types of data to obtain richer and more accurate point cloud measurement data, providing more reliable theoretical support for on-site rapid tree obstacle ranging and analysis.

[0103] Using 2D-2D, 3D-3D, 3D-2D data registration algorithms and multi-sensor clock synchronization technologies, high-precision fusion of lidar and visible light camera data is achieved, providing precise measurement data for tree obstacle ranging and analysis. Using data registration algorithms and clock synchronization technologies, high-precision registration of lidar and visible light camera data is achieved, data deviation caused by device time asynchronization and spatial position differences is eliminated, and the accuracy of data fusion is improved; Data fusion combines the advantages of the two types of data to obtain richer and more accurate point cloud measurement data, providing more reliable theoretical support for on-site rapid tree obstacle ranging and analysis.

[0104] The main components of the radar-vision intelligent device system module include the following parts:

[0105] Integrate multiple sensors to achieve comprehensive perception of the surrounding environment of the transmission line, integrate data from different sensors, and provide a comprehensive data basis for subsequent analysis;

[0106] Real-time data modeling and rapid ranging technology, which can construct models of the transmission line and the surrounding environment in real time, quickly calculate the distance between the tree and the transmission line, and realize real-time monitoring and analysis of tree obstacles;

[0107] Carry out intelligent ranging analysis of the safety distance, automatically judge whether the tree obstacle threatens the safety of the transmission line, improve the intelligence, real-time performance and accuracy of tree obstacle detection, and reduce the cost and risk of manual detection.

[0108] Integrate lidar, visible light cameras, POS sensors, etc., and use real-time data modeling and rapid ranging technology to achieve intelligent and real-time detection and analysis of tree obstacles on transmission lines, improving the detection accuracy and efficiency. Integrate multiple sensors to achieve comprehensive perception of the surrounding environment of the transmission line, integrate data from different sensors, and provide a comprehensive data basis for subsequent analysis; Real-time data modeling and rapid ranging technology, which can construct models of the transmission line and the surrounding environment in real time, quickly calculate the distance between the tree and the transmission line, and realize real-time monitoring and analysis of tree obstacles; Carry out intelligent ranging analysis of the safety distance, automatically judge whether the tree obstacle threatens the safety of the transmission line, improve the intelligence, real-time performance and accuracy of tree obstacle detection, and reduce the cost and risk of manual detection.

[0109] Specifically, first, based on the research on point cloud acquisition, modeling, compression, and transmission technologies, quasi-real-time point cloud data was output, providing a support basis for subsequent radar-vision fusion feature matching and rapid tree obstacle analysis;

[0110] Through point cloud acquisition, modeling, compression, and transmission, technologies such as the fast incremental triangulation method, the improvement of point cloud model reconstruction efficiency, the V-PCC dynamic point cloud compression method, and the efficient mapping and encoding of point cloud information were used to improve the efficiency of pre-scanning and post-processing of three-dimensional lidar point clouds. The proposed algorithms not only improved the efficiency and accuracy of point cloud modeling but also achieved efficient compression of dynamic point cloud data, providing new ideas and methods for the research on intelligent sensing modeling, compression, and transmission algorithms. At the same time, these innovative points have broad potential in practical applications and can promote technological progress and application expansion in related fields.

[0111] Secondly, through the fusion of multi-source perception information by accurately registering lidar and visible light cameras, theoretical support was provided for on-site rapid tree obstacle ranging and analysis;

[0112] In the radar-vision perception system, lidar and visible light cameras respectively provide lidar point clouds and video image data. These two types of data have different advantages and limitations. To obtain high-precision point cloud measurement data, accurately registering them is the key to data fusion processing. The registered data can complement each other, providing richer and more comprehensive information, which is beneficial to point cloud distance measurement. Using 2D-2D, 3D-3D, and 3D-2D data registration algorithms, combined with multi-sensor clock synchronization technology, high-precision inter-frame fusion matching of multi-source perception data was carried out. The theoretical accuracy that this data can reach is within 10 centimeters. At the same time, using the spatial coordinate system mapping and conversion algorithm, the data was synchronized to a unified coordinate system, combined with three-dimensional space measurement technology and multi-scale feature extraction technology, so as to obtain accurate three-dimensional space measurement data, providing a theoretical basis for issuing on-site real-time and rapid tree obstacle analysis reports.

[0113] An innovative radar-vision intelligent device system was developed, realizing the ranging and analysis of tree obstacles on transmission lines, and improving the intelligence, real-time performance, and accuracy of tree obstacle ranging;

[0114] Finally, using lidar three-dimensional space measurement and two-dimensional image target detection technologies, through multiple simulations and data fitting based on the actual transmission line site, a complete set of lidar-vision intelligent device systems for quickly generating tree obstacle measurement data on-site has been developed. As a comprehensive perception system, the lidar-vision fusion device integrates lidar, visible light cameras, and various POS sensors. Through a series of technical processes, it realizes precise environmental perception and data fusion, not only providing rich point cloud data for the client but also supporting in-depth data analysis to generate high-precision ranging data, providing strong data support for various application scenarios. In terms of data processing in the lidar-vision system, the device can collect a variety of information data, including three-dimensional laser point clouds, visible light video images, and device attitude data. Through the research and implementation of the lidar-vision fusion intelligent device, using real-time data modeling and rapid ranging technologies, intelligent ranging analysis based on the safe distance below the transmission line is carried out, improving the intelligent, convenient, real-time, efficient, and accurate measurement of the safe distance in the transmission line corridor.

[0115] A method for detecting tree obstacles on transmission lines based on lidar-vision fusion, the detection method includes the following steps:

[0116] S1. Data acquisition: Use lidar to scan the transmission line and its surroundings to obtain three-dimensional laser point clouds and determine the spatial position; collect visible light video images through a camera to capture the texture and color of the trees; use a POS device to obtain the real-time position and angle information of the device to ensure the spatial consistency of the data;

[0117] S2. Data preprocessing: Denoise and filter the lidar data to improve data accuracy; successively grayscale, normalize, and denoise the visual image to enhance the image quality for subsequent processing;

[0118] S3. Point cloud processing and transmission: Quickly collect three-dimensional point cloud data around the transmission line, use rapid incremental triangular meshing technology for modeling, and use the V-PCC method to compress the data to achieve quasi-real-time transmission, providing support for subsequent analysis;

[0119] S4. Feature extraction and matching: Extract the geometric features of trees and transmission lines from the lidar data, extract the texture and color visual features from the visual image, and then perform feature matching to improve the accuracy of tree obstacle detection;

[0120] S5. Multisensory information fusion: Use 2D-2D, 3D-3D, 3D-2D data registration algorithms and multi-sensor clock synchronization technologies to accurately register and fuse the lidar and visible light camera data, providing theoretical support for tree obstacle ranging and analysis;

[0121] S6. Tree obstacle detection and analysis, calculating the distance between trees and transmission lines, classifying tree obstacle levels based on tree species, size, and distance, analyzing historical data to predict the growth trend of trees, and judging the danger level of tree obstacles;

[0122] S7. Result output and early warning, organizing the detection data and analysis results to generate a report to provide data support for operation and maintenance; setting an early warning threshold, and when the tree obstacle reaches the danger threshold, promptly warning and alarming the operation and maintenance personnel to avoid accidents.

[0123] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting tree obstacles on transmission lines based on the fusion of radar and vision, characterized in that, include: The data acquisition module uses laser radar, cameras and POS devices to comprehensively collect the original data of the surrounding environment of the power transmission line, providing basic data support for subsequent links; The data preprocessing module removes noise and filters the lidar data, grays, normalizes and reduces noise on the visual images, improves data quality and creates conditions for feature extraction; The feature extraction and matching module extracts key features from the lidar and visual image data and establishes corresponding relationships, so that the two types of data can be mutually verified and provide information for accurate detection of tree obstacles; The tree barrier detection and analysis module calculates the distance between trees and transmission lines, classifies tree barriers according to tree type, size, and distance, and analyzes historical data to predict tree growth trends, providing a direct basis for preventive measures; The result output and warning module generates a report based on the test results, sets warning thresholds, and issues timely warnings when dangerous situations occur; Point cloud acquisition, modeling, compression and transmission module, which can quickly acquire point cloud data, use fast incremental triangulation technology to model and compress, and realize quasi-real-time transmission; The multi-sensing information fusion module uses 2D-2D, 3D-3D, 3D-2D data registration algorithms and multi-sensor clock synchronization technology to fuse lidar and visible light camera data to provide measurement data for tree obstacle ranging and analysis; The Leishi intelligent device system module integrates laser radar, visible light camera, and POS sensor, and uses real-time data modeling and fast ranging technology to achieve intelligent and real-time detection and analysis of tree obstacles on power transmission lines.

2. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, characterized in that: The data acquisition module mainly includes the following parts: LiDAR collects 3D laser point clouds, terrain undulations around transmission lines, tree heights and locations, and the spatial layout of towers and conductors, providing a geometric basis for subsequent distance measurement and spatial analysis; The camera collects visible light video images to clearly present the tree type, bark texture, leaf shape and color, helping to identify the growth characteristics of different tree species and assist in judging tree obstacle risks; The POS device obtains the device posture information, records the real-time position and angle of the device when collecting data, ensures the spatial consistency of the laser point cloud and video images, and enables subsequent data fusion and analysis to be performed in a unified coordinate system.

3. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, characterized in that: The data preprocessing module mainly includes the following parts: De-noising and filtering of LiDAR data to remove abnormal points caused by measurement errors and electromagnetic interference, and reduce data redundancy; Grayscale, normalization, and noise reduction of visual images. Grayscale is convenient for image feature extraction algorithms to process. Normalization makes images under different lighting conditions have a unified standard and enhances image comparability; Noise reduction removes noise from images, improves image clarity, and provides image data for subsequent image recognition and analysis.

4. The method for detecting tree obstacles in a transmission line based on radar-vision fusion according to claim 1, wherein: The feature extraction and matching module mainly consists of the following parts: Extract features from LiDAR data to obtain geometric features of trees and power lines, such as tree volume and power line curvature, to provide a quantitative basis for distance calculation and tree obstacle analysis; Extract features from visual images, capture the texture and color visual features of trees, provide intuitive visual information for tree species identification and tree obstacle classification, and supplement the deficiencies of lidar data in texture and color information; Feature matching, establish the spatial correspondence between lidar data and visual images, so that the two types of data can verify and supplement each other.

5. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, wherein: The main components of the tree obstacle detection and analysis module include the following parts: Calculate the distance between the tree and the transmission line, accurately measure the actual distance between the two, provide a quantitative standard for judging the danger level of the tree obstacle, and determine whether corresponding measures need to be taken; Identify and classify tree obstacles, and classify tree obstacles into different levels according to factors such as tree species, size, and distance; Analyze the growth trend of trees, predict the future growth direction and speed of trees through the analysis of historical data, and give early warnings of potential tree obstacle risks.

6. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, wherein: The main components of the result output and warning module include the following parts: Generate a tree obstacle detection report, record various data and analysis results of tree obstacle detection; Set the warning threshold and give warning alarms. When the tree obstacle distance or growth trend reaches the danger threshold, send warning and alarm signals to the operation and maintenance personnel in a timely manner.

7. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, characterized in that: The main components of the point cloud acquisition, modeling, compression and transmission module include the following parts: Point cloud acquisition, quickly and comprehensively obtain the three-dimensional point cloud data of the surrounding environment of the transmission line; Point cloud modeling, use the fast incremental triangulation technology to construct a point cloud data model diagram, and intuitively display the spatial relationship between the transmission line and the surrounding environment; Point cloud compression, adopt the V-PCC dynamic point cloud compression method to reduce the data volume without losing key information; Point cloud transmission, realize quasi-real-time point cloud data transmission, so that the operation and maintenance personnel can obtain the latest information on the surrounding environment of the transmission line in a timely manner.

8. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, wherein: The main components of the multi-sensor information fusion module include the following parts: Use the data registration algorithm and clock synchronization technology to realize the registration of lidar and visible light camera data, and eliminate the data deviation caused by the device time asynchronization and spatial position difference; Data fusion, complement the advantages of the two types of data to obtain point cloud measurement data.

9. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claim 1, characterized in that: The main components of the lidar-vision intelligent device system module include the following parts: Integrate multiple sensors to achieve all-round perception of the surrounding environment of the transmission line and integrate the data of different sensors; Real-time data modeling and fast ranging technology, real-time construct the model of the transmission line and the surrounding environment, and quickly calculate the distance between the tree and the transmission line; Carry out intelligent ranging analysis of the safe distance, and automatically judge whether the tree obstacle threatens the safety of the transmission line.

10. The method for detecting tree obstacles on transmission lines based on radar-vision fusion according to claims 1-9, characterized in that, This detection method includes the following steps: S1. Data acquisition, use lidar to scan the transmission line and its surroundings to obtain three-dimensional laser point cloud and determine the spatial position; collect visible light video images through the camera to capture the texture and color of the trees; obtain the real-time position and angle information of the device with the help of the POS device; S2. Data preprocessing, denoise and filter the lidar data; gray-scale, normalize and denoise the visual image in turn to enhance the image quality; S3. Point cloud processing and transmission, quickly collect the three-dimensional point cloud data of the surrounding area of the transmission line, model it with the fast incremental triangulation technology, and compress the data using the V-PCC method; S4. Feature extraction and matching: geometric features of trees and transmission lines are extracted from lidar data, and texture and color visual features are extracted from visual images, followed by feature matching; S5. Multi-sensor perception information fusion: 2D-2D, 3D-3D, and 3D-2D data registration algorithms and multi-sensor clock synchronization technology are used to register and fuse lidar and visible light camera data; S6. Tree obstacle detection and analysis: calculate the distance between trees and transmission lines, classify tree obstacle levels according to tree species, size, and distance, analyze historical data to predict tree growth trends, and judge the degree of danger of tree obstacles; S7. Result output and early warning: organize the detection data and analysis results to generate a report; set an early warning threshold, and when the tree obstacle reaches the danger threshold, promptly send an early warning to the operation and maintenance personnel.

Images