Research on Clouds Registration Algorithm Based on Multi-view Point
Jian Yang · Journal of Changchun University of Science and Technology · 2014
Aiming at the low automatic matching efficiency of point clouds, the stitching precision remains to be improved,in this paper,a stitching algorithm of automatic matching marked point is put forward. Firstly,according to the space invariant feature among marked point, a dynamic distance matrix is introduced to record the distance of marked points that are searched,marked points are matched by using cyclic iterative method to compare dynamic distance matrix. Finally,the multi-view point clouds are stitched by using the method of least squares to solve the coordinate transformation matrix.Through the experimental comparison,the point cloud registration algorithm is obviously superior to the traditional method,finally stitching precision is verified. Multi-view point clouds stitching experiments show that the method is convenient, fast and practical and suits for industrial production, and the stitching precision of0.0179mm can be achieved.