AN ALGORITHM OF PIPELINE POINT CLOUD EXTRACTION USING Li-MEDIAL SKELETON

摘要

In order to solve the problem of pipeline extraction in massive point cloud, we propose a new method for automatic segmentation of dense pipeline with circular section based on the algorithm of Li-medial skeleton construction. The algorithm mainly consists of four steps. First, the point cloud of plane is filtered out based on general algorithms such as RANSAC to obtain high quality data. Then, the point cloud is downsampled using octree decomposition and the Li-medial skeleton is calculated in parallel by the neighborhood point set of the sample points which is extracted using the Euclidean cluster extraction according to the spatial distribution of the point cloud. Finally, we reorient the skeleton to the cylinder axis along the direction of the skeleton, and analyze the connectivity of the skeleton to get the result. We apply this algorithm to scenarios such as simulation data and power plants with incomplete cloud data and dense pipelines. The experimental results show that the algorithm proposed in this paper can extract pipeline and obtain skeleton quickly and accurately when the data is incomplete.