The K-nearest neighbor fast searching algorithm of scattered data

Abstract

The paper has put forward an improved K-nearest searching algorithm of scattered data, which is significant to the technology of surface recreate of reverse engineering. Firstly, the initial segmentation of point cloud space is made by adopting the traditional block algorithm, and then estimates the average dot pitch of point cloud. Re-divide the point cloud space according to average dot pitch. The block result decreases the searching range of k-nearest neighbor searching algorithm.