Unmanned aerial vehicle point cloud and ground three-dimensional laser scanner point cloud registration method

Abstract

The invention discloses an unmanned aerial vehicle point cloud and ground three-dimensional laser scanner point cloud registration method. The method mainly comprises the steps: S1, acquiring the oblique photography point cloud and three-dimensional laser scanner point cloud; S2, preprocessing the point cloud data; S3, extracting and organizing feature point pairs of the two kinds of point clouds; S4, registering the model; S5, restoring parameters; S6, carrying out the point cloud overall registration. According to the method, the characteristics and rules of the two kinds of point clouds are analyzed, the common homonymy point pairs are selected, coarse parameter values are estimated, and a parameter correction matrix is solved, so the conversion matrix is obtained, and registration of the two kinds of point clouds with different sizes is achieved. The problems that the quality of a measured object top point cloud model obtained based on a three-dimensional laser scanner is poor, unmanned aerial vehicle oblique photography is good at obtaining the measured object top point cloud, and the top point cloud is poor are solved; meanwhile, the method can be qualified for point cloud registration with large angle deviation, and can solve the problem of proportionality coefficient of two kinds of point cloud sizes.

Description

technical field [0001] The invention relates to the registration of two different types of point clouds, in particular to a method for registration of a point cloud of a drone and a point cloud of a ground three-dimensional laser scanner. Background technique [0002] With the development of computer hardware, the advantages of point cloud data are reflected, and the methods of obtaining point cloud data are also different. There are 3D laser scanners, including terrestrial 3D laser scanners, airborne 3D laser scanners, and handheld 3D laser scanners. The 3D laser scanner can quickly obtain the point cloud model of the measured object, but when the terrestrial 3D laser scanner measures the object, the point cloud of the top and the self-occluded area will be missing due to the self-occlusion of the measured object. Oblique photography can also obtain the point cloud model of the measured object, and has its advantages in obtaining the point cloud on the top of the measured ...

AI Technical Summary

Problems solved by technology

The 3D laser scanner can quickly obtain the point cloud model of the measured object, but when the ground 3D laser scanner measures the object, the point cloud of the top and the self-occluded area will be missing due to the self-occlusion of the measured object.

The existing automatic registration methods are not suitable for point cloud data with a large number of missing point clouds, and the accuracy of the two point clouds is different from the coordinate system, so it is not only the problem of rotation and translation that needs to be considered

Images