A Method for Determining Normals of Large-Scale Dense Point Clouds

Abstract

A method for determining the normal of a large-scale dense point cloud, which solves the problem of determining the efficiency of the normal. The point cloud data are all considered to be on the same plane. For the cube containing the point cloud, an interpolation node attached to the surface is constructed to form a series of interpolation nodes. Compute the normals of all interpolated nodes by neighborhood search, principal component analysis. Project all point clouds in the same cube and all interpolation nodes in adjacent cubes to the coordinate plane along the direction of the maximum component of the normal to obtain a series of two-dimensional points, and use bilinear interpolation to find the normal of each two-dimensional point. Use the projection correspondence to assign the normal to the corresponding 3D point cloud data to complete the calculation of the point cloud normal. The beneficial effect is that the times of neighborhood search and normal line calculation are greatly reduced, normal line calculation results with different precisions can be obtained by setting the side length of the cube, and the method is convenient and flexible to use.

Description

Technical field: [0001] The invention belongs to the field of optical three-dimensional measurement data processing, and relates to a method for determining normals of large-scale dense point clouds. Background technique [0002] The development of optical three-dimensional measurement technology enables the acquisition of large-scale dense point clouds within a few seconds, and the model of an object often requires more than one million three-dimensional point clouds to express. Because the use of 3D point cloud data is very convenient, it has been widely used. The normal calculation of point cloud data is the basis of feature extraction, data reduction, data smoothing, surface reconstruction, etc., and is the key basic technology of point cloud data processing. [0003] Three-dimensional point cloud data is often scattered in space, and there is no direct connection between points. Therefore, the normal calculation of point cloud data usually first divides the scattered d...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problem of low normal calculation efficiency of large-scale dense point cloud data, the present invention regards the point cloud as a whole as a curved surface, adopts a space block processing strategy, and divides the point cloud data into very small space cubes, each cube All point cloud data can be considered on the same plane

Images