基于局部的全自动网格修复算法

An Automatic Surface-Based Mesh Repairing Algorithm

为了提高复杂三维外形模型的网格修复鲁棒性和准确性,提出一种基于局部的全自动网格修复算法.首先,对瑕疵类型进行自动识别,将其自动分划为洞、退化三角形、自交和非流形等不同类别;然后,针对不同瑕疵类别分别采用类动态规划方法、混合内核方法和裁剪缝合方法等相应的策略进行修复;最后,设计出一个完善的修复流程以确保高效、准确地完成整体的网格修复.以Thingi10K数据集为例,以修复后的模型能否被填充四面体为标准,与6种经典算法进行对比,实验结果表明,所提算法不仅可以准确地保持原模型的几何特征,并且其修复成功率为90%.

In order to improve the robustness and exactness of mesh repairing for complex 3D models,an automatic surface-based mesh repairing algorithm is presented.Firstly,the flaws are automatically identified and divided into different categories such as holes,degenerate triangles,self-intersection,nonmanifold,etc.Secondly,for different flaw categories,the appropriate strategies such as dynamic programming like method,hybrid kernel method,and cutting and stitching method are used to repair mesh accordingly.Finally,an appropriate process is designed to ensure the mesh repairing is effective and efficient.Taking whether the repaired model can be filled with tetrahedrons as a standard,the proposed approach is compared with the six state-of-the-art approaches on Thingi10K dataset.The experimental results show that the proposed algorithm can keep geometric features of the original models,and the repairing success rate is 90%.