Nearest neighbor point set quick search method based on Hash-Cube spatial hierarchy partition structure

Abstract

The invention provides a nearest neighbor point set quick search method based on a Hash-Cube spatial hierarchy partition structure. The method comprises the following steps: step 1, through estimating point cloud number included in unit length, dividing a whole point cloud space into a plurality of cubes in the same size, all points in the point cloud being wrapped in the cubes; step 2, through calculating cube coordinates of each point, combining with a direct addressing characteristic of a Hash function, distributing each data point to the corresponding cube; step 3, through sorting the data points, storing the data points, that is, each cube Hash-Cube just recording the address of the point whose search value is minimum; step 4, performing nearest neighbor query on spatial points, in a nearest neighbor search process, according to sample radius r of a current point and the Hash function, rapidly establishing a search domain, searching points in a search radius range in the search domain, to determine a nearest neighbor point set, and finally complete searching. Using the method can greatly save storage space and improve searching efficiency.

Description

technical field [0001] The invention relates to computer data and query technology, in particular to a fast retrieval method for the nearest neighbor point set based on the Hash-Cube space hierarchical division structure. Background technique [0002] The rapid development of three-dimensional measurement technology makes the scanning of entities more and more fine, and the scale of point clouds obtained is quite large. How to quickly and effectively find the nearest neighbor point set of any point from these scattered point cloud data has always been a problem. A relatively popular research direction, many point cloud-based digital geometry processing, such as point normal vector calculation, point cloud stitching, point cloud denoising, point cloud meshing, etc., all rely on efficient and feasible nearest neighbor query structure. [0003] At present, there are mainly three types of retrieval methods for the nearest neighbor point set, including brute force search, Voronoi...

AI Technical Summary

Problems solved by technology

[0009] With the rapid development of three-dimensional measurement technology, the physical image obtained by scanning is becoming clearer and the data volume is also increasing. The problem brought about by this is that the traditional search method for the huge data volume will take a lot of time. Therefore, Need a search method that is fast, consumes less resources, and has high usage efficiency

Images