Point cloud automatic registration method based on point feature histogram

Abstract

The invention discloses an automatic point cloud registration method based on point feature histograms, which includes: collecting and obtaining point cloud data from multiple stations of a measured object, determining a base station and a station to be registered; and randomly selecting m from the cloud data of both sites. sampling points, so that any three points are not collinear, and every two sampling points satisfy the distance constraint principle; match the two stations with the same point pairs, so that the base station sampling point set and the station to be registered sampling point set form Points with the same name are approximately congruent to the triangulation network; obtain the transformation matrix of the site cloud data to be registered, calculate the point feature histogram difference value to obtain the registration error; filter out the transformation matrix corresponding to the minimum registration error value; calculate the cloud data of the two sites The overlapping area point cloud data is used to obtain the fine registration matrix through the iterative closest point algorithm. The invention can solve the problem of easily falling into a local minimum solution when registering point cloud data of two adjacent sites when the cloud overlap is small or the features of the overlapping area are not obvious. The method has high accuracy and high overall efficiency.

Description

technical field [0001] The invention belongs to the field of point cloud data registration, in particular to an automatic point cloud registration method based on a point feature histogram. Background technique [0002] 3D laser scanning technology emerged in the reverse engineering of the industrial field in the early days, and was mainly used in the industrial manufacturing fields such as handicraft design, automobile, aircraft and shipbuilding, and was gradually applied to the surveying and mapping industry. Usually, the 3D laser scanning work is affected by factors such as the wide range of the measured object, the limited measurement range of the scanning equipment, and object occlusion. It is necessary to set up multiple stations to scan at different angles to obtain the surface of the measured object. Overall point cloud data. According to the scanning principle of the scanner, the collected 3D point cloud coordinate data is not in a unified coordinate system, so it ...

AI Technical Summary

Problems solved by technology

[0005] An object of the present invention is to provide a point cloud automatic registration method based on point feature histograms, which can solve the problem of point clouds in the prior art when the overlapping degree of adjacent two site clouds is small or the characteristics of the overlapping area are not obvious. It is easy to fall into the problem of local minimum solution during data registration, and this method has high accuracy and high overall efficiency

Images