基于机载 LiDAR 系统的电力线点云提取方法研究

针对野外地形高度起伏大、走向复杂、杆塔形式多样且点云密度不均匀等造成电力线点云提取精度低的问题，文章提出一种自动化精度高的新型电力线点云提取方法。根据电力线走向对整体点云进行分块处理；使用迭代高程阈值法计算得到最佳高程分割阈值，将地面点、地物点与电力线点进行分割；采用DBSCAN密度聚类法聚类出杆塔点并进行剔除，对于误分到杆塔点的部分电力线点结合点云高程搜索进一步提取出来；使用欧式聚类法对单根电力线点进行精提取。实际输电线路点云数据实验表明，该方法电力线总提取精度为99%，实现了电力线点云的自动、高效提取，可满足实际电力巡检生产需求。

Aiming at the problem of low extraction accuracy of power line point cloud caused by high fluctuation of field terrain height, complex trend, diverse forms of towers and uneven density of point cloud, a new power line point cloud extraction method with high automatic accuracy is proposed. The whole point cloud is segmented according to the power line direction. The iterative elevation threshold method is used to calculate the optimal threshold value of elevation segmentation, and the ground point, ground object point and power line point are segmented. DBSCAN density clustering method is used to cluster and eliminate the tower points, and part of the power line points mistakenly assigned to the tower points are further extracted by searching. Euclidean clustering method is used to extract single power line points. According to the actual transmission line point cloud data experiment, the total power line extraction accuracy of this method is 99%, the automatic and efficient extraction of power line point cloud is realized, and the actual demand of power inspection can be met.