A novel identification method based on point cloud data processing technology for quadric surface models

Abstract

A novel method is proposed in this paper to detect the types of quadric surface models. Firstly, using high-precision 3D scanner, point clouds are acquired from quadric surface models. Secondly, triangulated irregular network models are generated from the point clouds by Delaunay algorithm. Thirdly, normal vectors of the nearest neighborhood planes of points in point clouds are acquired by the least square fitting algorithm. Finally, mapping relationship between the distributions of normal vectors and the types of quadric surface models is constructed. As a result, in real applications, if the distribution of normal vectors obtained for a certain scanned quadric surface model is consistent with the constructed mapping relationship, surface type of the model can be rapidly identified. The proposed method is validated by using an experiment.