多尺度下车载激光点云数据中电力线的提取  

Combined Use of Geometric Features at Multi-Scales for Power-Line Extraction in Terrestrial Lidar Point Clouds

作　　者：贾魁 张晶[1] 

机构地区：[1]首都师范大学资源环境与旅游学院,北京100048

出　　处：《首都师范大学学报（自然科学版）》2018年第1期72-76,共5页Journal of Capital Normal University：Natural Science Edition

摘　　要：电力巡线工作是输电线路安全的保障,传统人工巡线的方式费时、费力.激光雷达技术可以快速获取电力线的真三维点云信息,可以通过提取电力线点云数据,最终服务于电力巡线工作.本研究首先使用Kdtree分析点云邻域内的点集合,并使用多个递增尺度分析电力线点云数据的几何结构特征并提取出各尺度下呈线分布特征的点云数据,最终叠加各尺度提取结果,比较单个尺度同合并各个尺度后的电力线点云数据提取结果的完整度,实验结果说明叠加多个尺度的提取结果完整度比各个尺度的提取完整度都高.Monitoring the power lines is of great importance to ensure the safety of transmission grid. The traditional mencentric way is time-consuming and laborious. The Li DAR technology has provided an effective way for acquiring accurate 3 D information of objects,and has been widely used in monitoring the power lines. This paper proposes a new approach to extract power lines in the Li DAR point clouds. Firstly,the geometric features of each Li DAR point are analyzed under multiple scales. Then the points of line structure are extracted at each scale. Furthermore,the extracted line structure with different scales is combined to generate a final extraction result. The accuracy assessment demonstrates that the multi-scale extraction result could achieve better completeness than those single scale extraction results.

关 键 词：多尺度 电力线提取 点云 几何结构特征 

分 类 号：P225.2[天文地球—大地测量学与测量工程]