基于运动特征的数字几何处理算法研究

项目名称： 基于运动特征的数字几何处理算法研究

项目编号： No.61272392

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 许威威

作者单位： 杭州师范大学

项目金额： 76万元

中文摘要： 传统的数字几何处理算法偏重于利用低层次的几何信息，如曲率等局部微分信息，没有充分考虑几何模型中所含的特征信息，使得用户难以控制算法以生成满足特定领域需求的处理结果。为了解决此问题，国内外前沿研究提出了将几何中所蕴含的高层次特征信息或特定领域知识集成到数字几何处理中，以得到高效，更符合用户要求的几何处理算法。 本项目通过对几何模型的动态特征，即其所能产生的运动，与其几何形状的关系进行深入研究，以期实现考虑运动特征的数字几何处理算法，使得用户能够在创建或编辑几何形状时能够有效满足其最终运动的需求。项目的研究内容包括考虑运动动特征的几何建模、三维重建、基于蒙特卡洛马尔科夫采样的几何优化和四边形网格化算法。

中文关键词： 数字几何处理；运动特征；蒙特卡洛过程采样；三维重建；快速仿真

英文摘要： Traditional digital geometry processing algorithms focus on low level geometry information, such as local curvature information. They do not fully consider the characteristic information implied in the 3D geometry models. It makes the user hard to control the algorithms to produce the models which can satisfy the requirements of specific domains. To solve this problem, the state-of-the-art geometry research emphasizes to integrate the high level geometry features implied in the geometry into the procedure of geometry processing, such as modeling, editing and 3D reconstruction, to improve the efficiency of the geometry processing algorithms. We propose to consider the relationship between the dynamic feature of the geometry models, i.e.the motion that they are capable of, and their shapes. We expect to develop geometry processing algorithms which are aware of the motion requirements so that the user can take control of the final motion in the modeling and editing of geometry models. The research content includes motion-aware geometry modeling,3D reconstruction, mesh animation capture, Monte Carlo Markov Chain sampling based geometry optimization and quadrangle mesh generation.

英文关键词： Digitial Geometry Processing；Motion；Mote-Carlo Sampling；3D Reconstrunction；Fast Simulation