Aiming at the problems of low registration accuracy and high computational complexity of existing algorithms for point clouds,a point cloud registration algorithm based on multi-feature fusion was proposed.The normal angle of the point cloud,the projection distance,curvature and Euclidean dis-tance variance of the point and its neighborhood points were extracted,and the feature points of the point cloud to be registered extracted by fusion.The iterative closest point algorithm based on the Gaussian probability model was used to register the feature point set,so as to realize the accurate registration of the noise point cloud.In the experiment,Cup and Bunny public point cloud data and cultural relic point cloud data were used to verify the proposed registration algorithm,and the results showed that the accura-cy of the algorithm was about 20％higher than that of the existing algorithm,and the time consumption reduced by about 25％.
point cloud registrationmulti-feature fusionnormal anglecurvatureprobability iterative closest point
维普
万方数据
