基于三角域的自适应细分曲面的构建

发布时间：2018-03-23 05:04

本文选题：Loop细分　切入点：细分曲面　出处：《西北农林科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：细分曲面技术具有传统参数曲面、隐式曲面造型技术不具备的诸多优点,因此广泛应用于计算机辅助几何设计和计算机动画造型等领域。但细分曲面的所有细分模式都存在一个亟待解决的问题:对模型进行细分时,在每一层都是全局细分,随着细分次数的增多,网格的面片数成指数级增长,巨大的数据量使得细分后的模型难以进行其它处理。然而实际情况下并不需要对整个模型细分来获得光滑的曲面,仅需在不平坦或曲率较高的区域进行细分使得这部分区域更光滑。此外对一块已光滑的区域实施细分迭代操作,对改善模型的光滑度效果很不明显。针对这个问题,本文利用控制网格的局部信息,提出了一种基于二面角阈值和Loop模式的自适应细分算法,利用该算法可避免在相对光滑处再细分,与正常细分相比,既大大减少了数据量,提高了模型的处理速度,又达到了对模型进行细分的目的。本论文的主要工作围绕下面三点展开:(1)对细分方法的基本理论和方法进行了比较全面的阐述,并对细分曲线算法、DooSabin细分算法、Catmull-Clark细分曲面算法、蝶型细分模式、√3细分模式等几种典型细分模式学习并实现。(2)对目前现有的自适应算法进行比较,分析了他们的优点和不足。针对Loop细分模式实现了基于二面角阈值的自适应细分模式。该算法构造了新的细分规则,实验结果表明该方法在保证曲面质量的同时,控制了细分面片数量的过快增长。(3)在ubuntu 16.04桌面版64位系统下,使用C++语言和OpenGL、CGAL、GLUT技术,采用可以快速查询网格信息的半边型数据结构和方便读写的off模型文件,初步实现了自适应Loop细分曲面造型算法。同时,该算法可以实现对模型的旋转、缩放等操作。并选择龙、正四面体、兔子模型进行了全局细分和自适应细分,进行对比分析。结果表明,针对不同的模型,面片减少率9.4%-82.0%,从而节省了大量的CPU和显存资源。
[Abstract]:Subdivision surface technology has many advantages, such as traditional parametric surface, implicit surface modeling technology, Therefore, it is widely used in computer aided geometric design and computer animation modeling, but all subdivision models of subdivision surface have a problem to be solved urgently: when subdividing the model, it is a global subdivision in each layer. With the increase of the number of subdivision times, the number of mesh patches increases exponentially, and the huge amount of data makes it difficult for the subdivided model to be processed otherwise. However, in practice, it is not necessary to subdivide the whole model to obtain smooth surfaces. It is only necessary to subdivide the region with uneven or high curvature to make it smoother. In addition, the effect of subdivision iterative operation on a smooth region is not obvious to improve the smoothness of the model. In this paper, an adaptive subdivision algorithm based on dihedral angle threshold and Loop mode is proposed based on the local information of the control grid. The algorithm can avoid subdivision at a relatively smooth point, which greatly reduces the amount of data compared with normal subdivision. It improves the processing speed of the model and achieves the purpose of subdividing the model. The main work of this paper is to expound the basic theory and method of the subdivision method around the following three points. Several typical subdivision models, such as the Doo Sabin subdivision algorithm Catmull-Clark subdivision surface algorithm, butterfly subdivision model and 3 subdivision mode, are studied and implemented. Their advantages and disadvantages are analyzed. The adaptive subdivision pattern based on dihedral angle threshold is implemented for Loop subdivision mode. A new subdivision rule is constructed. The experimental results show that the method not only guarantees the surface quality, but also guarantees the surface quality. This paper controls the rapid growth of the number of subdivision facets.) under the ubuntu 16.04 desktop version 64 bit system, using C language and OpenGL language CGALALLUT technology, adopting the half-edge data structure which can query the grid information quickly and the off model file which is easy to read and write. At the same time, the algorithm can realize the rotation, scaling and other operations of the model, and select the dragon, tetrahedron and rabbit models for global subdivision and adaptive subdivision. The results show that, for different models, the chip reduction rate is 9.4-82.0, thus saving a lot of CPU and memory resources.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.7

【相似文献】