Study On The Estimation Method Of Vector 3D Point Cloud Method

In recent years, with the rapid development of the 3D point cloud acquisition equipment, low cost of acquiring the point cloud data of the object to be measured more details of the possible, related computer graphics research based on point cloud is also influenced by academic circles and industry, which also contributed to the computer graphics point cloud based on the Renaissance [1] An attribute vector as the essential method of point cloud data, the drawing method and surface reconstruction algorithm of point cloud based on plays an important role, especially in the 3D reconstruction algorithm, the degree of accuracy of normal vectors directly affects the results of three-dimensional reconstruction based on spline function, such as B [3] and other implicit surface reconstruction algorithm based on radial basis function [2],will completely depend on reliable and accurate method of vector. Therefore, to obtain accurate, reliable method of vector is very necessary. However, due to the noise is unavoidable in the scanning device, image and depth discontinuities in the depth image, directly through the scanning method to obtain the vector obtained is difficult to satisfy the requirements of drawing and surface reconstruction. Therefore, is particularly important in the unorganized 3D point clouds is the normal vector estimation method.This paper first introduces the domestic and international researching situation on estimation of 3D point cloud method and the existing main research method. Then it introduces the method of 3D point cloud normal vector used in this paper and the preliminary estimation of 3D point cloud normal vector redirection method, finally through the improved algorithm, the realization of vector 3D point cloud method to estimate the system and carries on the analysis comparison to experimental data obtained. By using the method of vector method for 3D point cloud preliminary estimation in this paper is based on the local plane fitting method. The method is assumed for complex point cloud systems, local area of any point is smooth, it can be a good fit for a plane, therefore can obtain the normal vector of each 3D point. For the normal vector redirection 3D point cloud is adopted by the improved method of plane fitting method based on vector redirection, based on the original method, increase the normal vector direction of propagation constraints and local redundant edge constraints, also from the original starting point an initial spread into multiple propagation starting point. Finally, according to the improved method, realized the vector 3D point cloud method to estimate the system, through the concrete experimental data show that the improved method not only in the running efficiency of the program have very big promotion, and get more accurate method of vector.