Research on Optimization of Point Cloud Registration ICP Algorithm

Point cloud data, as a basis for the three-dimensional data types has attracted more and more attention these years. compared with other types of three-dimensional data, point cloud data can be collected by simple ways, and it contains surface texture, surface color, and other types of features on the surface of the target object. Recently, in the field of cloud data processing, many new point cloud processing methods have been proposed on the basis of existed theory, among which, the most notable one is the newly proposed algorithm in the field of point cloud registration. In this paper, an improved point cloud registration algorithm based on classical ICP algorithm is proposed. In the classical ICP algorithm, the main part which limit the efficiency of the algorithm is the iterative search process of the corresponding point, the new algorithm proposed in this paper uses the 4-point coincidence algorithm to accelerate the corresponding point search process as an improvement to the classical ICP algorithm, and to verify its improvement by experiment.