Abstract: An improved method is summarized based on quadtree and grid to sort the point cloud data
in this paper. Using least squares method fitting projection plane, the 3D point cloud data become
two-dimensional after projection. Envelope edges are calculated through convex hull. And the
envelope regions are gridded. An improved quadtree indexing method is used to associat the data
point with the grids. The completion of sort depends on these grids. Finally, using MATLAB to
achieve example codes demonstrating the feasibility of the proposed method.

Key words: point cloud, quadtree, grid association, data organization