Research On The Simplification And Registration Algorithm Of Three-dimensional Point Cloud Data

With the development of three dimensional laser scanning,it has become an important method to obtain the 3D information of a certain object rapidly due to its outstanding advantages and characteristics of initiative,high precision,high real-time and easy expression on information and data collection.3D point cloud data is presented in the three-dimensional coordinates of points,and the general data format has high practical value in deformation analysis,model measurement,digital display,virtual restoration,virtual reality,target reconstruction.When simplifying and registering the data model,traditional grid method and curvature method have some shortcomings,such as the feature points are easy to be lost,the boundary is not complete,and the traditional ICP algorithm requires a high point cloud overlap.Based on what are inferred above,the K-means clustering algorithm based on boundary retention and point cloud registration algorithm based on the consistent sphere are proposed.For the simplification algorithm,firstly,the k-d tree is used to initialize the centroid,and then X-Y boundary extraction algorithm is used to preserve the integrity of the boundary,finally,the clusters are subdivided according to its curvature,which will retain the necessary number of points in the high curvature region and retain some uniform distribution points in flat regions.And the superiority of the method is verified with experimental results.For the registration algorithm,on finding the corresponding points,the proposed algorithm innovatively combines the rotation invariance of the sphere with the SVD orthogonal consistency algorithm based on neighborhood,which could ensure that registration algorithm can obtain the corresponding points with higher accuracy,and on this basis,spread and get more corresponding points,and then use the rigid constraints to eliminate the error points,finally use quaternion method to solve transformation matrix.The results show that the proposed algorithm has not only overcome the shortcomings of the traditional ICP algorithm,but has better accuracy and can be applied to practical engineering.