Fast registration method of 3D point cloud data with color information

Abstract

The invention discloses a fast registration method of 3D point cloud data with color information. The method comprises a step of counting color values in a target point set and a registration point set and the quantity distribution of points with a same color and filtering the color of RGB point cloud, a step of screening points with a same RGB value in two point sets and then setting a certain color tolerance for the registration point set, a step of extracting four initial points from one point selected in the target point set in each time, constructing a candidate point set similar to the characteristics of the four initial points in the registration point set, and judging whether the candidate point set is same name points of the four initial points, a step of allowing the four initial points and the corresponding four same name points to form four point pairs, using a total least square algorithm to calculate a rigid transform matrix, and completing the global registration of the registration point set, and a step of using a neighboring point iteration method to complete a final local registration finally. According to the method of the invention, the efficiency of point cloud registration can be greatly improved under the premise of maintaining high precision, and the registration time is shortened.

Description

technical field [0001] The invention belongs to the technical field of three-dimensional (3D) laser point cloud data processing, relates to the field of 3D point cloud registration application technology, in particular to a fast global registration technology specially used for 3D RGB point cloud data. Background technique [0002] With the rapid development of laser scanning equipment, the acquisition of 3D laser point cloud data has become easier and easier, and the research on 3D point cloud data processing has become more extensive, such as: 3D point cloud registration, 3D object recognition, 3D Modeling etc. Among them, 3D point cloud registration is to splice 3D point cloud data measured multiple times into a complete data set under the same coordinate system, and since point cloud registration is the first step in point cloud data processing, the accuracy of the registration results It directly affects the follow-up work such as point cloud segmentation, feature extr...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the existing 3D point cloud registration technology, there are mainly the following problems: 1) The traditional 3D point cloud registration steps mainly include: key point extraction, feature descriptor calculation, search strategy for corresponding point pairs, and ICP fine registration, These steps are slightly too many, making point cloud registration too complicated

Images