A Fast Weighted Registration Method of 3D Point Cloud Based on Curvature Feature

Abstract

In order to realize the fast and accurate registration of 3D point cloud data, a new fast weighted registration method is proposed in this paper. Firstly, using curvature feature, the method samples the original 3D point cloud data to quickly find matching points and remove wrong point pairs. Secondly, by introducing the iterative re-weighted least squares (IRLS) algorithm, the method carries out coarse alignment of the scattered point cloud. Finally, the method presents an improved distance-weighted Iterative Closest Point (ICP) algorithm to achieve fine matching. The experimental results show that the method has good convergence, robustness and accuracy.