An interactive approach to the point cloud based pipeline construction

Abstract

The point cloud data obtained through the laser-scanner is very large in density with a lot of noise and outlier, so that the manual reconstruction of a pipelinemodeling is very time consuming and additional costs. Piping system especially, account for the greatest proportion of plant equipment. In this work, we have developed a system which can reconstruct the pipeline automatically based on the sequential segmentation and skeleton extraction.The developed approach consists of following steps. An initial step is,pre-process thepoint cloud with filtering and downsampling technique based on aPoint Cloud Library (PCL).The second step is segmentation of point cloud based on smoothness constraint. Next step is skeletonization of segmented data by medial axis extraction. The last step is,automatically generate pipelines to the skeletonized data.The presented results enables the reliable 3D models of pipelines, which could be incorporated into the reconstruction of a plant information modeling method and utilized for, assist maintenance and expansion of existing plants. © Research India Publications.