A Fast Registration Method for Massive Point Clouds Based on 3D-SIFT and 4PCS
The registration of measurement point cloud and model point cloud is the key of visual positioning.Aiming at the prob-lems of poor visual positioning accuracy and low algorithm efficiency caused by large amount of measurement point cloud data and low overlap rate with CAD model point cloud,a registration method of measurement point cloud and model point cloud based on the fusion of 3D scale invariant feature transform(3D-SIFT)and four point fast robust matching algorithm(4PCS)is pro-posed.Firstly,the depth camera is used to extract the point cloud of the part,and the extracted measurement point cloud is de-noised and filtered;Then 3D-SIFT feature point extraction algorithm is used to extract feature points from measurement point cloud and CAD model point cloud;Finally,the extracted feature points are used as the initial values of the 4PCS algorithm to achieve the registration of the two point cloud data.Compared with the commonly used 4PCS algorithm and Super-4PCS algo-rithm,the algorithm simulation and experimental results show that the proposed algorithm can improve the registration speed by more than 30%on the premise of ensuring the registration accuracy.
measurement point cloudmodel point cloudSIFT4PCSpoint cloud registration
国家科技期刊平台
NSTL
万方数据
