三角网格模型的特征保持混合折叠简化  

Feature preserving mesh simplification based on hybrid collapse

作　　者：曹增欢 黄常标[1] 郑红 CAO Zeng-huan;HUANG Chang-biao;ZHENG hong(Fujian Key Laboratory of Special Energy Manufacturing, Xiamen Key Laboratory of Digital Vision Measurement, Huaqiao University, Xiamen 361021, China)

机构地区：[1]华侨大学福建省特种能场制造重点实验室厦门市数字化视觉测量重点实验室,福建厦门361021

出　　处：《光学精密工程》2019年第4期971-983,共13页Optics and Precision Engineering

基　　金：国家科技支撑计划资助项目(No.2015BAF24B00);福建省科技重大专项资助项目(No.2014HZ0004-3);福建省引导性项目资助(No.2018H0020;No.2017H0019;No.2016H0020)

摘　　要：在增材制造、逆向工程等领域,广泛存在包含大量甚至海量数据的三角网格模型。为便于存储并提高处理效率,经常需要进行网格模型简化。但在网格简化过程中存在特征保持、简化率和简化效率冲突的问题,为更好地平衡简化结果和简化效率,提出了特征保持的混合折叠算法。在基于曲度精确计算新顶点、最大距离高效计算折叠代价的基础上,对边界特征区域和非边界特征区域采用边折叠方式进行保特征简化,对非特征区域则采用三角形折叠方法进行高效简化,最后通过偏差和网格正则度对简化结果作出误差评价。算法实例表明:混合折叠算法的模型细节特征保持较好,简化前后变形较小且效率适中。In additive manufacturing, reverse engineering, and other fields, many triangular mesh models were used to process big data. To facilitate storage and improve processing efficiency, simplifying the mesh model was necessary. However, conflicts existed among feature preservation, simplification rate, and simplification efficiency in mesh simplification. For improved balancing of results and efficiency of simplification, a method based on curvedness was proposed to compute new vertices more precisely. In addition, a method based on the max distance was proposed to compute collapse cost efficiently. The edge collapse was adopted for preserving the features of boundary and non-boundary feature regions, and the triangle collapse was applied to simplify the non-feature areas efficiently. The mesh regularity of the simplified model and the deviation between the original and simplified models were used as evaluation criteria for mesh simplification. Extensive experiments demonstrate that the mesh simplification algorithm with hybrid collapse can effectively preserve the features of a complicated mesh model with fewer errors and moderate efficiency.

关 键 词：三角网格模型 网格简化 混合折叠 细节特征保持 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]