The initial assignment of the two parameters, the original neighborhood radius rEps and density threshold pMinPts, leads to uncertainty in the extraction results of power line point clouds. On the basis of density clustering, an adaptive discrimination method for point cloud cluster classes has been added. This method avoids the tedious process of repeated testing of initial parameters by personnel, and the C++language is used to complete the development and testing of the algorithm for accurate extraction and fitting of power lines. The results show that the improved density clustering method has a loss rate of only 0.02％ in extracting power line point clouds, and a residual of 0.213 m in 3D reconstruction; this method greatly improves the accuracy and convenience of extracting power line point clouds and is suitable for power inspection and 3D reconstruction in high-voltage power corridors.
万方数据
维普
