Small point cloud matching based on bi-direction random KD-tree
F.W. Zhu, Shunke Lv, Zheng-Xiong Yuan · Fifth Asia International Symposium on Mechatronics (AISM 2015) · 2015
The Matching of the point clouds is a key technique in the three-dimensional tracking and 3D reconstruction process, as large amount of feature point data (point cloud) collected from different view-point on an object surface are required. An improved ICP algorithm based on bi-direction random KD tree and camera calibration for small point clouds pair was introduced in the paper. The experiment showed that the registration accuracy could be improved.