Abstract: In order to make the mesh smoothing algorithms preserve sharp features and optimize meshes, a feature-preserving uniform mesh smoothing algorithm was presented. Firstly, the presented algorithm combined local predictions from the second order triangular meshes with modified bilateral filtering and quasi-Laplacian smoothing operator, and then combined the advantages of Laplacian smoothing and mean curvature smoothing, the vertex was regulated in the direction of combination of the Laplacian and mean curvature methods vectors. Compared with the existing methods, experimental results showed that the presented algorithm could effectively remove noise from meshes while preserving the features of the original model and optimizing meshes.

Key words: reverse engineering, mesh smoothing, bilateral filtering, feature-preserving, uniform mesh

摘要： 为了使网格光顺算法在消除噪声的同时，保持模型特征并优化网格，提出一种保特征的网格均匀化光顺算法。首先获取二阶邻域三角面片的信息，使用改进的双边滤波器和准拉普拉斯算子预测顶点位置，然后综合拉普拉斯光顺法与平均曲率法的优点，沿顶点切平面和法矢量方向同时对网格顶点进行调整。实验结果表明，与现有方法相比，该方法可以有效去除噪声，保持三维表面模型的特征并使网格得到优化。

关键词: 逆向工程, 网格光顺, 双边滤波, 特征保持, 均匀网格