Research Of A Smoothing Algorithm Based On The Main Direction

Free-form curves and surfaces are widely used in shape design and reverse engineering. In CAD system, parametric curves and surfaces are usually applied to interpolate, approximate and fit the giving data. Nevertheless, As error exists while measuring and calculating data, it is necessary to do fairing operations on the curves and surfaces obtained by these data, thus to get the aesthetic curves and surfaces. Fairing curves and surfaces is becoming one of the most important problems in the research of computer-aided design.Although piecewise polynomial patches are still the dominating free-form surface representation in engineering applications, the use of triangle meshes has become increasingly important—especially with the rising complexity of 3D models. A computer graphics object reconstructed from real-world data contains undesirable noise and small-scale oscillation. An important problem is how to remove the noise and oscillations while preserving desirable geometric features of the object.The paper presents a set of mesh smoothing tools developed to increase mesh regularity, reduce over-smoothing, and enhance crease lines. Two of the most popular approaches for smoothing and de-noising polygonal surfaces are minimizing energy functional associated with differential-geometric surface characteristics and Laplacian smoothing. In this paper, we discuss Laplacian smoothing and modified methods, mean curvature flow, PDE method and the smoothing method based on directions. Smooth geometry models by different algorithm, analyses the smoothing surfaces and compare the efficiency. We hope to smooth mesh surfaces in less time and preserve desirable geometric features of the object.We present a smoothing algorithm based on the main direction combined with mean curvature flow and normal filter method. Our method preserves desirable geometric features and prevent from over-smoothing. It's linear and fast.