A registration method of point cloud to CAD model based on edge matching

Abstract Rough registration is foundational for accurate registration. Features are always used to construct correspondences between point cloud data and CAD model. Considering that edges are more common than feature points or feature planes. A registration method of point cloud to CAD model based on edge matching is proposed. The method includes three steps. Firstly, edge points (EPs) are extracted from point cloud and CAD model respectively. The extraction of point cloud data includes two steps, namely extraction of section line data and identification EPs. Then two groups of EPs are uniformly sampled with same interval and matched according to the deviation of direction angle sequence of EPs. Finally, based on the matching result of EPs, rotation angles of point cloud data to CAD model are obtained. Verification experiments are carried out on three edges on a template. Experiment results show that deviations of rotation angles are within 1 degree for all three edges. The proposed method can be widely used for the universality of edges on a part.