Research On Power Line Extraction Method Based On Vehicle LiDAR Data

The continuous innovation of science and technology has made the economic and social development of human beings more and more subject to environmental constraints and resource constraints.It is imperative to seek a new energy utilization model,the global energy Internet came into being that is relying on the growing information technology.As a key task of global energy Internet construction,the smart grid combines technologies such as the Internet,the Internet of Things,and intelligent mobile terminals to meet the diverse needs of society and serve the development of smart homes,smart communities,intelligent transportation,and smart cities.In the process of the global energy Internet and the smart grid construction,digitalization of power equipment,secure transmission of power networks and rational planning are essential.Traditional power network inspections are difficult to adapt to the development and safe operation of intelligent power grids due to their high operational intensity,long cycle and low efficiency.As an efficient remote sensing method,vehicle LiDAR technology can quickly acquire a large number of point cloud data with precise three-dimensional spatial coordinates in the target scene in a short time,which can be used to acquire the physical state of the power grid,realize the digitization of the power grid and the inspection of the power grid,thus providing assistance for the construction of the smart grid and the global energy Internet.Based on a comprehensive analysis of vehicle LiDAR system and the characteristics of point cloud data,this paper first studies the pre-processing process of vehicle LiDAR point cloud data,and then uses the Normal Distribution Transformation(NDT)algorithm to realize the rough extraction of power line points,and on this basis,uses the RANdom SAmple Consensus(RANSAC)algorithm to achieve the fine extraction of power line points.Finally,the classified power line points are analyzed and evaluated,the effectiveness and feasibility of this method are verified by several experiments.The main research contents are as follows:Firstly,based on the research goal of this paper is the power line,the relative height information between the target points and the ground points is the basis of elevation filtering.Due to the excessive terrain fluctuations in the survey area,the original data does not reflect the target points elevation information very well.In order to avoid this situation,in the point cloud data preprocessing stage,in addition to noise points culling and sparse points separation,ground points filtering is also required,and the relative elevation of the non-ground points is obtained according to the ground points information,the elevation filtering can be performed accordingly.Secondly,the current research on NDT algorithm mainly focuses on the point cloud registration direction in the field of Simultaneous Localization And Mapping(SLAM).There are few studies on its application to point cloud data extraction.This paper innovatively applies it to the extraction of power line.By dividing the point cloud data into multiple NDT units,and extracting the NDT units with linear distribution of point cloud data,the rough extraction of power line points is realized.On the basis of rough extraction of power lines,automatic segmentation of data is realized by determining the location of power towers.Then,the RANSAC algorithm is used to fit the straight line and the parabola in the horizontal and vertical planes respectively,so as to realize the accurate extraction of the power line points.Finally,the extraction results of point cloud data are qualitatively and quantitatively analyzed and evaluated,and the vehicle LiDAR data with different characteristics are selected for multiple experiments.The experimental results are compared and analyzed to verify the validity and feasibility of the research methods in this paper.The research methods are summarized to illustrate the inadequacies of this paper and the direction for further improvement.The analysis of the results shows that the proposed method can quickly extract the power line points in the vehicle LiDAR data,especially in the case of huge amount of vehicle point cloud data,the data processing efficiency advantage is obvious.The classified power line points have high precision and can meet the application requirements of subsequent 3D modeling and safety analysis.