摘要
Abstract
Due to the irregularity of point cloud data,there may be certain errors in feature extraction of tree barrier point clouds,which degrades the classification accuracies of ground and non-ground laser point clouds and further increases the errors of tree bar-rier recognitions in transmission line corridors.Therefore,an online identification method for tree barriers in transmission line cor-ridors based on airborne LiDAR is proposed in this paper.Firstly,aiming at various abnormal problems existing in the laser point clouds of transmission line corridors,appropriate algorithms are adopted for data correction and processing,and the spatial homoge-nization thinning method is used to simplify the laser point cloud data.Secondly,the least square method is introduced to fit the lo-cal quadratic surface of point clouds.Based on the curvature features of quadratic surfaces and local multi-scale features,complete power line information is accurately extracted from laser point clouds.Then,the ground slope separation algorithm is applied to distinguish ground and non-ground laser point clouds.On this basis,the Euclidean clustering algorithm based on K-D tree search is used to identify and separate vegetation point clouds from non-ground point clouds.Finally,a distance threshold for tree obstacle recognition is set to achieve online recognition of tree barriers in transmission line corridors.The experimental results show that the proposed method has excellent comprehensive performance in online recognition of tree barriers in transmission line corridors,and can effectively identify and mark potential hidden tree barriers.关键词
机载激光雷达/输电线路通道/树障/在线识别/二次平面拟合Key words
airborne LiDAR/transmission line corridors/tree barriers/online identification/quadratic plane fitting分类
信息技术与安全科学
