3D point cloud cluster analysis based on principal component analysis of normal-vectors

Technical demands for extraction of significant components from spatial models are increasing as 3D sensors and their application technology has been developed and popularized. In this paper, we propose the 3D point cloud cluster analysis based on the principal component analysis(PCA) of normal-vectors. The distribution of normal vectors depends on a 3D surface shape within the local neighborhood. We discussed the PCA of the distribution of normal vectors to the point cloud. The results of the experiment show that our method could classify a local point cloud as a plane, a boundary and a vertex.