Simplification of point cloud data for large-scale ellipsoidal complex surface

Abstract

Aiming at the issue that the current point cloud data simplification algorithm cannot accurately retain the subtle features of ellipsoidal complex surface, simplification algorithm of a point cloud data feature preservation based on principle component analysis (PCA) and adaptive mean shift (AMS) methods is proposed. First, K-means spatial clustering is used to make preliminary clustering and simplification of point cloud data. Second, the representative value of the subtle features of the point cloud data is calculated based on the PCA method. Then, according to the proposed improved adaptive mean shift method, the simplification of the point cloud data is developed. Finally, the point cloud data obtained by the denoising and smoothing of a large spherical crown workpiece collected at a certain equipment manufacturing site is used as the experimental object, and a comparative experiment is conducted between the proposed simplification algorithm and the k-means clustering algorithm, which proves the proposed algorithm can effectively maintain the subtle feature points of the ellipsoidal point cloud data.