计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (16): 176-182.DOI: 10.3778/j.issn.1002-8331.1907-0145

• 图形图像处理 • 上一篇    下一篇

综合曲率约束的在线网格模型分割方法研究

赵云成,王琳,盛步云,赵飞宇   

  1. 1.武汉理工大学 机电工程学院,武汉 430070
    2.湖北省数字制造重点实验室,武汉 430070
  • 出版日期:2020-08-15 发布日期:2020-08-11

Online Mesh Model Segmentation Method with Curvature Constraints

ZHAO Yuncheng, WANG Lin, SHENG Buyun, ZHAO Feiyu   

  1. 1.School of Mechanical and Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China
    2.Hubei Digital Manufacturing Key Laboratory, Wuhan 430070, China
  • Online:2020-08-15 Published:2020-08-11

摘要:

针对现有三角网格模型块分割方法普遍存在计算复杂度高,无法体现工程意义,综合效果不理想,不满足Web环境下高效快速分割等问题,提出一种面向Web环境的简单高效的三角网格模型分割方法。根据高斯曲率和平均曲率特性划分出网格模型的凹区域,在凹区域中依据最小负曲率阈值提取凹特征区域,结合区域中心特征线提取方法以及边界线闭合和优化算法构造出闭合分割线,通过分割线将三角网格模型分割成有意义的分块。依托开源数字几何处理软件MeshLabJS,运用WebGL的几何处理及图形渲染功能,在普林斯顿标准数据集和COSEG形状数据集上进行算法测试,验证所提方法能够在Web环境下快速、高效、有意义地分割三维模型。

关键词: 三角网格模型, 离散曲率, 网格分割

Abstract:

Aiming at the problems of high computational complexity and unsatisfactory for fast segmentation of triangular mesh model blocks in Web environment, a simple and efficient triangular mesh model segmentation method for Web environment is proposed. Firstly, the concave region of the mesh model is divided according to the characteristics of Gauss curvature and mean curvature. Then, the concave feature region is extracted according to the minimum negative curvature threshold in the concave region. Combined with the method of extracting the central feature line of the region and the algorithm of closing and optimizing the boundary line, a closed segmentation line is constructed to segment the triangular mesh model. Relying on the open source digital geometry processing software MeshLabJS and using the geometry processing and graphics rendering functions of WebGL, the algorithm is tested on Princeton standard data set and COSEG shape data set. The results show that the proposed method can segment three-dimensional models quickly, efficiently and meaningfully in the Web environment.

Key words: triangular mesh model, discrete curvature, mesh segmentation