摘要
传统网格生长法对孔洞数量庞大且孔洞类型复杂的三维网格模型修复效果不佳。针对该问题,将泊松方程应用于三角网格模型的孔洞修补。利用原始模型信息建立泊松方程,对输入模型曲面进行全局拟合,根据孔洞信息裁剪拟合得到的预测曲面并与原始孔洞模型缝合,通过孔洞边界区域法向量信息调整修补曲面的三角面片方向,达到特征增强的目的。实验结果表明,该算法对于结构复杂的多孔洞三维模型修补效果较好,对噪声鲁棒性强,在保留模型原始信息的同时能够准确还原孔洞区域特征。
The previous hole-filling algorithms cannot effectively repair the complex 3D mesh models with large number of diverse holes. Aiming at this problem,this paper applies the Poisson equation into hole filling of triangular mesh model. First of all,the Poisson equation is established based on the initial model information, and the input model surface is globally fitted. Then ,the fitting prediction surface is cut out according to the hole information and stitched to the original hole model. Finally, triangular patches of repaired surface are adjusted according to the normal vectors of hole boundary region method, so as to enhance features of the repaired model. Experimental results show that the proposed algorithm has better integrality and noise immunicty when applied to the complex model with a large number of diverse holes. It can restore the features of the holes area while preserving the initial information of the model.
出处
《计算机工程》
CAS
CSCD
北大核心
2017年第10期209-215,221,共8页
Computer Engineering
基金
国家自然科学基金"基于全局优化的破损兵马俑虚拟方法复原研究"(61373117)
高等学校博士学科点专项科研基金"破损兵马俑虚拟复原拼接关键技术研究"(20136101110019)
关键词
三角网格
孔洞修补
泊松方程
特征增强
隐式曲面
三角剖分
triangular mesh
hole-filling
Poisson equation
feature enhancing
implicit surface
triangulation
