A 4PCS Coarse Registration Algorithm Based on ISS Feature Points

In view of the problem of poor effect and low accuracy of point cloud registration with low overlap rate under different frames, this paper analyzes the extraction results of ISS, SIFT3D and Harris3D commonly used key points and proposes a rough point cloud registration method based on ISS feature point descriptor combined with 4PCS. Firstly, ISS algorithm is used to extract feature point descriptors from the source and target point clouds to simplify the huge number of original point clouds. Secondly, the 4PCS registration algorithm is used to find the corresponding points on the feature point set and calculate the corresponding transformation matrix to complete the coarse registration. The Stanford point cloud dataset tests show that: 1) for the point clusters with different overlap rates, the key points extracted by the ISS key point detector are more balanced and not affected by outliers; 2) The registration result of the algorithm in this paper is better than 4PCS in the case of low overlap rate and noise.