CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleCurve-Skeleton Extraction of 3D Model
Thesis Advisor张晓鹏
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline计算机应用技术
Keyword曲线骨架提取 形状分解 距离变换 Laplace光滑算子 Curve-skeleton Extraction Shape Decomposition Distance Transformation Laplacian Smoothing
Abstract随着计算机图形学的快速发展,三维几何模型被广泛地应用在计算机辅助设计、虚拟现实环境、数据可视化、医学影像、教育与娱乐等各个领域。曲线骨架作为三维形状的一维抽象表示方式,描述了三维模型几何形状特征和拓扑结构特征。三维模型的曲线骨架提取提供了对三维模型的形状和结构的高级理解与分析,在虚拟导航、模型简化、模型匹配与检索、形变和动画等方面有着重要应用。 本文首先概述了三维模型骨架提取的研究现状,对现有方法进行了分类与分析。针对体素数据和网格数据这两种常见的三维物体表示形式,文本进行了以下的研究: 1.提出了基于距离场的体素分解方法。对几何形状的根点和末梢点进行种子点距离变换,根据根点或末梢点出发的区域生长策略,自动判断分叉并优化分割面,将对象分解成有意义的部件且部件之间分界合理。 2.提出了结构化骨架提取方法。提取体素部件骨架并连接成对象的结构化骨架。在各部件中单独提取骨架避免了相邻分支在分叉处对骨架的影响,提高了骨架的中心性。结构化的整体骨架保持了原始模型的拓扑结构,骨架不会产生断裂和多余的杂枝。 3.基于骨架进行真实植物建模。证明算法能够处理真实的扫描数据,具有抗噪声的能力且重建准确度高。另外,该体素骨架化方法适用于带有环状结构的形体,能处理表面体素模型以及实心体素模型。 4.采用了基于Laplace算子的网格收缩方法。算法应用全局位置约束的隐式Laplace光滑算子,将网格几何收缩到体积近似为零的形状。收缩的网格可以通过连接性简化转换成一维的曲线骨架,并根据骨架和网格的映射关系修正骨架的中心性。收缩过程保留原始网格的关键特征,且基于迭代的隐式光滑操作对数据中的噪声不敏感。
Other Abstract3D models are widely used in many disciplines including computer aided design, virtual reality environment, scientific visualization, medical imaging, education and entertainment. Curve-skeleton is a thinned 1D representation of a 3D object. It abstracts the essential topology and geometry of the underlying object, which provides a higher level understanding of the shape and its structure. There are many important applications of curve-skeleton in the field of virtual navigation, reduced-model formulation, model matching, registration, morphing and animation. In this thesis, we present the state of the art in the research of curve-skeleton extraction. Many of the existing algorithms are categorized into classes and discussed about their performance. We do research of the skeleton extraction methods both on volume models and mesh models. The main work is as follows: 1.We propose an efficient method for shape decomposition in volumes based on the distance transform from root point or tip points. Accompanied by the region growing, junctions are automatically detected and boundaries between different components are produced reasonably. 2.We propose a structural skeleton extraction method of the volume. By connecting the sub-skeleton of the components, the whole skeleton keeps the topology of the original model and is highly centered. No redundant and disconnected branches are resulted in the skeleton. The skeleton extraction method works well with shapes with loop structures and both surface voxel models and volumes. 3.We apply the extracted skeletons into real plant reconstruction. The experiments show that our plant modeling method based on curve-skeletons has a precise reconstruction and it is not sensitive to the noise in the scanned data. 4.We present a simple and robust skeleton extraction method on mesh contraction. The method first contracts the mesh geometry into a zero-volume skeletal shape by applying implicit Laplacian smoothing with global positional constraints. The resulted skeleton preserves the shape of the original topology and is insensitive to noise.
Other Identifier200628014629087
Document Type学位论文
Recommended Citation
GB/T 7714
项波. 三维模型的曲线骨架提取[D]. 中国科学院自动化研究所. 中国科学院研究生院,2009.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20062801462908(3870KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[项波]'s Articles
Baidu academic
Similar articles in Baidu academic
[项波]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[项波]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.