基于区域离散曲率的三维网格分水岭分割  

Watershed 3D mesh segmentation based on regional discrete curvature

作　　者：贾晖[1,2] 耿国华[1] 周明全[3] 张建刚[4] 

机构地区：[1]西北大学信息科学与技术学院,西安710127 [2]西安邮电大学计算机学院,西安710121 [3]北京师范大学信息科学与技术学院,北京100875 [4]西安热工研究院有限公司电站信息及监控技术部,西安710032

出　　处：《计算机工程与应用》2015年第11期182-186,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61373117);西安邮电大学青年基金(No.ZL2013-28)

摘　　要：针对离散曲率估计对噪声敏感且特征值计算量大的特点提出了基于区域离散曲率的三维网格分水岭分割算法。寻找三维模型显著特征点;对三维模型进行预分割,确定分割带;在分割带区域上计算离散曲度极值点,利用测地距离和曲度极值点对三维模型进行分水岭分割。算法在分割前无需进行网格去噪,实验结果证明,对主体分支明显的模型具有较高的分割边缘准确度和较快的分割速度。The discrete curvature is sensitive to the noise and heavy computation of feature value. In order to segment the3 D meshes, this paper presents a novel and efficient watershed algorithm which is based on regional discrete curvature. It extracts prominent feature points, pre-segments and then obtains the segment region. It estimates the discrete curvature within the segment region. It utilizes the Gauss curvature as the feature watershed to segment the 3D mesh. This algorithm does not need to wipe off the noise at initial stage, the experimental results show that the proposed algorithm can accurately and frequently yield meaningful borderline which has an evident branch structure.

关 键 词：三维模型 网格分割 区域离散曲率 分水岭算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]