An Effective Approach for Rock Mass Discontinuity Extraction Based on Terrestrial LiDAR Scanning 3D Point Clouds.

This paper presents an effective approach for rock mass discontinuity extraction based on the terrestrial LiDAR scanning 3-D point clouds. First, a Quadtree-Octree index method that retained the adjacency relationship of rock point clouds was proposed to organize point cloud data in a high-efficiency manner. Second, the normal vector (NV) calculation was directly conducted by the local triangulated irregular networks together with eight neighborhood areas. Third, a double-clustering strategy was developed based on the Quadtree-Octree index and NV calculation discussed in Step one and two. The first clustering was conducted in the point cloud matrix at each station, whereas the second clustering among multiple stations. Fourth, an extended random sample consensus algorithm was designed with two separate checks (distance and angle) to detect whether the compliance with certain constraints occurred between rock point clouds and fitting planes. The proposed method were evaluated by a real field data set in China and a public data set from Rockbench. The feasibility of the proposed method were verified by these two data sets, which indicated a promising perspective for the field engineering survey.