Automatic Classification And Tree Barrier Early Warning Analysis Approach For Transmission Line Corridor From Airborne LiDAR Point Clouds

For a long time,traditional method for the power line inspection is carried out with manual labor or by helicopter carrying optical camera,to monitor the conditions of power line,which not only brings high cost and lower efficiency,but also is not accurate because of the limited ability of identifying objects of human eyes and the low positioning accuracy of the image.Laser scanner system(Li DAR)can obtain high precision three-dimensional geometric information of transmission line corridor.Using the Li DAR data can make it possible to effectively and automatically detect the tree risk point of the corridor.In addition,airborne inspection method can further improve efficiency and reduce costs.Therefore,the application of airborne Li DAR technology in the transmission line corridor inspection has wider application prospects.Based on 3D point cloud data obtained by airborne Li DAR,this paper classified the corridor point cloud according to the characteristics of point cloud data and different features of objects in the corridor.After that,the transmission line simulation in different conditions are carried out combined with meteorological conditions in the corridor and wire parameter information.On this basis,the early warning analysis of dangerous tree points of transmission line is completed.The related studies are as follows:First of all,this paper introduces the components and working principle of the airborne Li DAR system.And the data characteristics,storage forms and the common spatial index organization form of point cloud data is introduced in detail.We find a point cloud spatial index method which can effectively improve the efficiency of point cloud data feature extraction.Secondly,the power line classification based on curvature and multi-scale features in multi-dimension is proposed in this paper.According to the linear characteristics of power lines,the validity of curvature characteristics to identify power line points is discussed.Because the curvature estimation needs to construct a quadric surface for a point near point set,and the linear power line cannot complete this construction,the power line classification is performed accordingly.Then,combined with the multi-scale feature in multi-scale dimension,the noise points are removed and the complete power line points are preserved.Then,point cloud can be divided into ground points and non-ground points by cloth terrain filtering.Further,the pylon is located in the non-ground point by combining the priori pylon coordinates information provided by power line operator,and the pylon points are classified according to the cylindrical model.Since most objects of the corridors are vegetation,the non-ground points preserved after the pylon classification are classified as vegetation points.Finally,on the basis of the theory of overhead power lines,combined with meteorological conditions and wire parameters,the transmission lines simulation in different conditions are carried out.By simulating transmission lines of different conditions(such as high temperature,ice coating or wind),tree barrier detection is carried out using the power lines with maximum sag as the auxiliary data of point cloud,to realize the early risk warning analysis of transmission corridor.