虚拟现实系统中的可见光成像仿真  

Simulation of Lens System Imaging in Virtual Reality System

作　　者：王滨[1] 戴树岭[1] 

机构地区：[1]北航大学先进仿真技术重点实验室,北京100083

出　　处：《系统仿真学报》2007年第6期1312-1315,1319,共5页Journal of System Simulation

摘　　要：对虚拟现实系统来说,画面的逼真程度是其能否为使用者提供良好沉浸感的关键。传统的计算机小孔成像模型因缺乏必要的深度信息,不能满足医学、军事等领域的仿真需求。结合真实透镜成像系统的光学原理,对原有计算机成像模型进行了一定的补充和修改,提出透镜成像系统数学模型,并具体解释了如何将透镜系统中的焦距、光圈、弥散圈、景深等概念映射到计算机成像过程当中。随着科学技术不断发展,可编程图形硬件的出现使得大量的图形计算从CPU转移至GPU成为可能。介绍了一种对透镜成像系统数学模型的实现方法,它充分利用图形处理器(GPU)的多纹理技术和可编程性,不仅获得更真实的图像效果,且满足实时性要求,可以在虚拟现实系统中应用。For Virtual Reality （VR） system, one of the key elements for better user immersion is the fidelity of computer generate images. Because of the insufficient cue of depth, the traditional pin-hole camera model cannot be used in some areas, such as medical and military simulation. By combining real lens imaging system, a lens image system model was proposed, which is an amendment and modification to previous models, How to map some concepts of lens system was expatiated, such as the focal, the aperture, the confusion circle and the depth of field into the computer image system, With the development in science and technology, the programmable Graphics Processing Unit（GPU） makes it possible for transferring the considerable calculation from CPU to GPU. An implementation method for lens imaging system model was proposed, which makes full use of GPU＇s multi-texture and programmable ability. Both vivid images and real time requirement can be achieved. The proposed method can be used in VR systems.

关 键 词：虚拟现实系统 透镜成像系统数学模型 实时性 可编程图形处理器 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]