A 3D Point Cloud Registration Algorithm based on Feature Points

Abstract

The standard ICP algorithm is a mainstream of 3D point cloud registration technique, which can establish effectively correct corresponding point set.An improved ICP algorithm based on inflection point is proposed.Our algorithm doesn't require the measured object having obvious features.As feature point, inflection point is easy to be obtained.Because not all of the data points are involved with, the amount of corresponding points is reduced.The experiment results show that our algorithm is more effective than other similar algorithms.Especially for large-scale data points, the algorithm has more advantages.