Tree point clouds registration using an improved ICP algorithm based on kd-tree

Abstract

The light detection and ranging (LiDAR) technology plays an important role in obtaining the three-dimensional information. A large number of point cloud data of the objects can be obtained through the LiDAR technology. The Iterative Closest Point (ICP) algorithm was widely used for registering the point cloud data, which typically only scan an object from one direction at a time. However, massive point cloud data has brought a great number of troubles to this registration method. The k-d tree is similar to the general tree structure and it can store, manage and search data efficiently. Therefore, an improved ICP algorithm which based on k-d tree was presented for tree point cloud data registration in this paper. The results showed that the improved ICP algorithm can improve the speed of registration about 10 times higher, and it also has obvious advantages in accuracy of registration.