Ordered Parameterization for Scattered Point Clouds

Abstract

In the numerical prediction process of large-scale spacecraft de-orbiting reentry disintegration in fall around the end of life after the service, the key foundation of automatic reconfiguration for disassembled objects is to form new disintegrated shape in real time. The patch point cloud parameterization is an important part of geometric reconstruction to arrange scattered point clouds so as to facilitate surface fitting and other needs. The random scattered sheet-like point cloud has no regularity, and it is difficult to fit a smooth surface. It is necessary to carry out ordered parameterization on such point cloud to facilitate surface entity construction or regular mesh division. A Projection Vector Weighted (PVW) method for scattered point clouds is firstly presented for curved patches with small curvature and smooth point cloud boundary. The method comprises the following steps of: firstly, establishing a base surface according to the boundary of a scattered point cloud; secondly, the point cloud is projected onto the surface to obtain projection coordinates and vectors; then, discretizing the base surface, and searching neighborhood points of a discrete point in the projection point cloud; and finally, carrying out inverse distance weighting on these vectors corresponding to the neighbor points to correct the coordinate of the discrete point, so as to obtain a ordered point of the curved surface where the scattered point cloud is located. The experiments are carried out on the scattered point clouds, and the ordered parameterized point clouds are successfully obtained, with an average accuracy of 99.87% of the geometric maximum size. The method is suitable for the random scattered point cloud with small curvature, and easy to implement and efficient without iterative calculation to achieve higher precision under the condition of excellent based surface establishment.

Keywords: scattered point cloud, ordered parameterization, based surface, projection, inverse distance weighting