A Approach To Models Of Three-dimensional Grid Smoothing

In the computer graphics and the CAD, the three-dimensional entity is common presented by polygonal mesh .With the progress of technology in three-dimension scanning and curved surface rexonstruction development of technology,we can get these models surface polygonal mashes conveniently very,but the surface includes the noise.In design field ,in fitting of disorder poihts,in smooth surface and projecting of lamination,there is a great demand for the smooth curved surface.So,the smoothing research of the three-dimensional meshes model has already become an important direction in which the three dimensional model has dealt with.In this paper, We lucubrates the most popular approaches for smoothing and denoising of polyhedral surface. One is Laplacian smoothing. It is a simple and fast method. So far the common technique for mesh smoothing, it moves recursively each vertex of the mesh by a displacement equal to the average of the neighboring vertices, but it will produces shrinkage. The other kind of method is Curvature smoothing algorithm which is proposed by Desbrun,etc. It moves every vertex in the normal direction with speed equal to a discrete approximation of the mean curvature at the vertex. Taubin introduced knowledge that signal deal with Laplacian smoothing, propose to alternate two scale factors of opposite signs with negative factor of larger magnitude in a weighted Laplacian smoothing flow. Such smoothing does not produce shrinkage. In addition, Liu XinGuo, Karbacher, Kuriyama, etc. have improved these two kinds of main algorithms.This paper is based by the studys mentioned above, used VTK(Visualization Toolkit)and C ++ programming,realized these most basic algorithms of Laplacian smoothing, Taubin smoothing and Curvature flow. Proceed with three respects of primitive mechanism, efficiency and error rate have compared and analyzed the advantage and disadvantage of each kind of algorithm, use a large number of experimental datas to prove from two of and. Finally, merge the result analysed and proved, a new improved method in surface smoothing is presented here which allowed to choose different smoothing algorithms based on the defferent attribute vertices of the mesh, solve the drawback that only used a kind of algorithm to appear in the past, hope to reach one kind balancedly in two respects of smoothing efficiency and smoothing result.It is proved that [the kind of improvement is effiective by experiment.