Title: Iteratively weighted principal component analysis and orientation consistency for normal estimation in point cloud

Authors: Bo Wen; Bo Tao; Wei Pan; Guozhang Jiang

Addresses: Key Laboratory of Metallurgical Equipment and Control Technology, Ministry of Education, Wuhan University of Science and Technology Hubei, China ' Key Laboratory of Metallurgical Equipment and Control Technology, Ministry of Education, Wuhan University of Science and Technology Hubei, China ' School of Mechanical and Automotive Engineering, South China University of Technology Guangdong, China; Department of R&D, OPT Machine Vision Tech Co., Ltd Guangdong, China ' Hubei Key Laboratory of Mechanical Transmission and Manufacturing Engineering, Wuhan University of Science and Technology Hubei, China

Abstract: In this paper, we present a method to robustly estimate normal of unorganised point clouds, namely Iterative Weighted Principal Component Analysis (IWPCA). Since the neighbourhood of a point in a smooth region can be well approximated by a plane, the classical Principal Component Analysis (PCA) is a widely used approach for normal estimation. Iterations are applied and bilateral spatial normal weights are introduced in each iteration for the local plane fitting to enhance the reliability of the PCA results. Minimal Spanning Tree (MST) is used to reorient flipped normals. We demonstrate the effectiveness and robustness of the proposed method on a variety of examples.

Keywords: point cloud normal; local plane fitting; least squares; iterative weighting; principal component analysis; minimal spanning tree; orientation consistency.

DOI: 10.1504/IJWMC.2020.111213

International Journal of Wireless and Mobile Computing, 2020 Vol.19 No.3, pp.267 - 275

Received: 25 May 2020
Accepted: 23 Jun 2020

Published online: 30 Oct 2020 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article