A New K Nearest Neighbours Algorithm Using Cell Grids for 3D Scattered Point Cloud

  • Jianhui Zhao Wuhan University
  • Chengjiang Long Wuhan University
  • Shuping Xiong Ulsan National Institute of Science and Technology
  • Cheng Liu Wuhan University
  • Zhiyong Yuan Wuhan University
Keywords: K nearest neighbours, search algorithm, 3D point cloud, cell grids

Abstract

In this paper, we propose one novel and efficient K nearest neighbours search algorithm based on 3D uniform cell grids. First, a simple min-max box is used to store all the points and a twice division strategy is adopted to determine the edge length of basic grids. Then we limit the search space for each grid with certain query points to control the amount of distance calculations under a suitable level. And the computational cost of sorting operations is reduced effectively through avoiding the unnecessary calculations, with the help of properly determined subspaces and sphere spaces for points. Compared with existing related algorithms, our method can search for the K nearest neighbours accurately and quickly, and it has many possible applications in the fields using 3D scattered point clouds.

DOI: http://dx.doi.org/10.5755/j01.eee.20.1.3926

Published
2013-12-30
Section
SYSTEM ENGINEERING, COMPUTER TECHNOLOGY