An improved method for registration of point cloud

A new hybrid least square method was proposed for registration of point cloud in this paper. The registration process was accomplished through two steps: the coarse registration and the accurate registration. The point cloud was transformed to the vicinity of the 3-D shapes by using genetic algorithm during the coarse registration procedure and the accuracy of point cloud registration was strongly raised in the accurate registration stage with iterative closest point algorithm. Experimental results show that the registration rate, matching accuracy and convergence rate of the algorithm proposed here are all improved. (C) 2017 Published by Elsevier GmbH.