大型无人机电力巡检LiDAR点云安全距离诊断方法

Safety Distance Diagnosis of Large Scale Transmission Line Corridor Inspection Based on Li DAR Point Cloud Collected With UAV

输电线路走廊中地物安全距离检测是电力运维部门日常线路巡检作业中的重要一环。结合此项运维实际需求,提出一种无人机电力巡检LiDAR点云数据的自动安全距离诊断方法:首先依据已知的线路位置与走向对无人机巡检系统采集的LiDAR点云进行裁剪,获得线路走廊区域点云;其次,采用自适应分区滤波的方法滤除地面点,获得非地面点云;在此基础上,在非地面点数据中,依据点云维数特征以及空间几何分布特性从非地面点中分离出导线点、杆塔点、及林木植被、建筑物等安全距离诊断所关注的线路走廊其他地物;继而提出一种迭代最小二乘电力线悬链线模型解算方法,自无序电力线激光点云数据拟合悬链线方程,用于后续安全距离计算;最后使用分段剖面安全距离计算方法计算线路与线路走廊下方地面/地物距离,并与标准安全距离做比较,对于距离小于安全距离的区域进行危险预警。采用大型无人机电力线路巡检系统采集的多组LiDAR点云数据对文中算法进行实验验证,并对检测结果与人工点云量测值与实地巡检值进行了定性与定量的对比分析。实验结果表明提出的安全距离诊断方法能准确探测树障等安全距离超限地物。

Safety distance diagnosis of ground objects in transmission line corridors is an important part in transmission line inspection task of power operation and maintenance departments. Combined with actual needs of power line maintenance, in this paper, an automatic safety distance diagnosis method of large scale transmission lines corridor inspection based on LiDAR point cloud collected with unmanned aerial vehicle （UAV） is proposed. Firstly, LiDAR point cloud collected with UAV inspection system are trimmed according to known power line position and direction, thus the point clouds in line corridor area are obtained. Secondly, adaptive partition filter is used to screen out ground points to obtain non-ground point clouds. Thirdly, power lines, towers, forest vegetation, buildings and other features of ＂safety distance diagnosis concem are separated from the non-ground points according to characteristics of point-cloud dimension and spatial geometric distribution. On basis of aforementioned procedures, an iterative least squares catenary model for power lines is proposed to fit catenary equation of power line laser point cloud data for subsequent safety distance calculation. Finally, a sectional safety distance calculation method is used to calculate distance between power line and ground objects in the line corridor and compare it with standard safety distance,then a danger warning is issued for the area with less distance than safety. The proposed method is validated with multiple LiDAR point cloud data collected with UAV power line inspection system. The results are compared against manual measurements in point clouds and on-spot field inspection to apply comparative analysis both qualitatively and quantitatively Experimental results show that the safety distance diagnosis method proposed in this paper can effectively detect over-limit objects such as tree barricades with safety distance measurement accuracy at decimeter level at UAV inspection flying height.