【作者】 苏诺；

【导师】 季桂树；

【作者基本信息】 中南大学 ， 计算机科学与技术， 2009， 硕士

【摘要】 实时碰撞检测是虚拟现实中一个非常关键的问题,其基本任务是确定两个或多个物体彼此之间是否发生接触、接触面积大小和穿透的深度。尽管针对碰撞检测已有了大量有价值的研究成果,但随着人们对交互实时性、场景真实性要求的不断提高,碰撞检测技术所面临的问题也日益突出,其中最核心的问题是如何有效地提高碰撞检测的速度。本文对碰撞检测相关技术进行了深入的研究,主要包括以下几个方面的内容:首先,从图形硬件发展的历史开始,介绍和分析最新GPU在通用计算方面的应用及其技术原理和发展状况。然后,对目前现有的碰撞检测算法进行了分类归纳,同时总结了一般碰撞检测算法所采用的总体框架,并着重介绍与分析了基于层次包围体树和基于图像空间的碰撞检测算法。在此基础上,本文提出了一种基于GPU的对参数化表面的碰撞检测方法。通过使用几何图像表示的参数化表面,可以实时的生成GPU优化的包围体层次结构,然后在这个层次结构的基础上实现优化的基于GPU的层次碰撞检测算法。结果显示本方法可以有效的提高碰撞检测的速度。更多还原

【Abstract】 Real time collision detection is one of the most important problems in the fields of virtual reality. Its fundamental task is to detect whether there are contacts or penetrations between two or among multiple objects. While, there have been many research achievements on solving the problem of collision detection, this problem is better to be solved with the more high demands of real time interactivity and realistic simulation of motions of virtual objects thereafter. This paper studies the technique of collision detection deeply. It contains the following parts:Firstly, starting from a brief introduction to some historical events on graphics hardware development, a detail introduction and analysis will be given to the technique and the latest development of GPU for general purpose computations.Secondly, this thesis reviews recent researches on collision detection and summarizes the general framework of a collision detection algorithm. Then it focuses on introducing and analyzing the collision detection algorithms based on bounding volume hierarchy and image space.Finally, base on the above contents, the thesis describe a GPU-based collision detection method for parameterized surfaces by geometry images allows to generate GPU-optimized bounding volume hierarchies in real-time that serve as a basis for an optimized GPU-based hierarchical collision detection algorithm. The experimental results show that our method is faster than the CPU-based method.更多还原