激光点云中电力线的提取和建模方法研究

Power line extraction and modeling based on LiDAR point cloud

当利用LiDAR激光点云进行电力线路危险点检测时,需要对电力线进行建模。为此提出了一种在激光点云中提取电力线并建模的算法。该算法首先利用激光雷达回波信息和电力线在三维空间中的线性特征,过滤掉大部分非电力线的点,然后在点云的XY平面内利用霍夫变换提取最长且相互平行的线性地物作为电力线,连接因数据遮蔽而造成的段线,并利用线的交点和沿线小网格内高程变化特征找出杆塔位置,最后在电力线断面内对多根电力线进行分割,分别计算悬链线模型。基于云南某500 kV线路实际数据的实验表明,该方法自动化程度高,电力线提取和建模准确度高。针对进一步提高该算法的通用性和自动化程度,提出了下一步的研究方向。

In the danger point detection of power line using LiDAR point cloud,the power modeling is needed.An algorithm for transmission line extraction and modeling based on point cloud is proposed.In the algorithm,most non-power line points are eliminated by LiDAR waveform information and linear features of power line in 3D space in the pre-processing.The points are projected onto XY plane and the power lines are detected by Hough transformation to extract the longest and parallel landmarks,and the short power lines due to data gap are linked.Tower position is detected by the power line intersection points and height variation features in adjacent 3D grids along the power lines.The last step is splitting power lines in the power lines section,and respectively calculating the catenary model.The experiment result of a 500kV power line in Yunann Province demonstrates the high-level automation and accuracy of this method.