Multi-view point cloud registration method based on K-means clustering center local curved surface projection

Abstract

The invention discloses a multi-view point cloud registration method based on K-means clustering center local curved surface projection. The method includes: giving an initial global transformation matrix; calculating a multi-scale feature descriptor and a normal vector of each frame of point cloud; determining clustering attribution of each point in a complete point cloud; calculating a multi-scale feature descriptor and a normal vector to obtain a registration corresponding point of the original point relative to the complete point cloud; carrying out bidirectional interpolation projection on the local MLS curved surface, and if a rigid body transformation consistency constraint condition is not met, eliminating the point pair, and obtaining a final matching point set of the single-frame point cloud; if the rigid body transformation consistency constraint condition is met, taking the projection point and the corresponding point thereof as a correct corresponding point pair; registering the N view point clouds in sequence; and achieving global optimization. According to the invention, the problem that the point cloud resolution is reduced and the registration precision is not high due to the down-sampling operation of the laser point cloud with sparsity originally is solved, that is, the sampling sparsity of the three-dimensional laser point cloud is solved.

Description

technical field [0001] The invention belongs to the field of three-dimensional point cloud reconstruction, in particular to a multi-view point cloud registration method based on local curved surface projection of K-means clustering center. Background technique [0002] Point cloud registration is an important step in 3D reconstruction technology, and its registration accuracy directly affects the results of 3D reconstruction. Although the dual-view registration process can accurately register two point clouds, there must be non-overlapping areas between the two point clouds, and registration errors are inevitable. The error will continue to increase, so multi-view global optimization is required. [0003] Some methods use low-rank sparse matrix factorization for multi-view registration in order to take advantage of the single-closed-loop constraint of multi-view point clouds. For scanning point clouds with a large overlap rate, the relative motion information redundancy in...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a multi-view point cloud registration method based on the local surface projection of the K-means clustering center, which performs a downsampling operation on the sparse laser point cloud, which leads to a decrease in point cloud resolution and lower registration accuracy. Not a high problem, that is, the sampling sparsity of the 3D laser point cloud

Images