关键词: 三角网格, 曲率计算, Gauss曲率, 平均曲率, NURBS曲面拟合

Abstract:

This paper proposed a new algorithm to estimate Gauss and mean curvature on triangular meshes based on NURBS fitting. In this algorithm, double neighbors were selected for local fitting in the first place which involved border checking. For parameterization, directly projection method was used. Then by using least square NURBS fitting, surface function was acquired, which followed the traditional curvature estimation method. Through tests on meshes considering different resolution and noises, final conclusion can be drawn that our algorithm has good results and is more stable than other algorithms, thus is more applicable.

Key words: triangular mesh, curvature estimation, Gauss curvature, mean curvature, NURBS fitting