中文摘要：

针对由点云数据重建得到的多边形网格模型存在孔洞的问题,提出了一种保持尖锐特征的多边形网格模型孔洞修补算法。首先,利用基于径向基函数的插补方法获得能够近似地逼近孔洞区域的光滑隐式曲面。然后,利用正则化匹配算法对该隐式曲面进行三角化,完成孔洞填充曲面片与原始模型的孔洞边界的缝合。最后,针对孔洞边界存在尖锐特征进行特征增强处理。实验结果证明,该算法对多边形网格模型的孔洞修补取得良好效果,并能够有效地恢复孔洞区域包含的尖锐特征。

英文摘要：

Due to existing holes in polygon models which is reconstructed from the digitized points,a sharpness dependent hole filling approach of polygon models is proposed in this paper.First,interpolation based on radial basis function is applied to create a smooth implicit surface and this surface can approximate the shape of the missed data.Then,a regularized marching tetrahedral algorithm is adopted to triangulate the above implicit surface.A repaired mesh model is obtained by stitching the hole patch and the hole boundary of the original model.Finally,a feature enhancement process is applied if there exists a sharp feature on the hole boundary of the original model.Experiment results show that our approach can produce excellent reparation results especially for recovering sharp feature.