基于GPU的烟雾数据体面混合绘制技术研究  

A GPU Based Hybrid Volume and Polygon Rendering Algorithm for Smoke Data

作　　者：徐华兵[1,2] 朱登明[2] 邱显杰[2] 刘金刚[1] 

机构地区：[1]首都师范大学计算机科学联合研究院,北京100037 [2]中国科学院计算技术研究所,北京100190

出　　处：《计算机仿真》2010年第2期212-217,共6页Computer Simulation

基　　金：国家自然科学基金项目(60603082);国家科技支撑计划(2008BAI50B07)

摘　　要：研究烟雾等大规模体数据在虚拟场景中实时可视化的技术是一种新的体面混合绘制技术,在灾害现象仿真中有着重要的应用。在处理体面混合绘制时,已有的绘制算法难以实时地处理体数据的遮挡和裁剪,并在内窥绘制中出现片元空洞。针对以上问题,提出一种基于GPU的体面混合绘制算法,将体数据视为参与介质,同时将面绘制的结果作为体绘制的积分域约束条件,从而把面绘制中的裁减算法推广到体数据的绘制中,实现了体面混合绘制的无缝融合。仿真结果表明算法能够实时地绘制虚拟场景中的大规模体数据,并且能够处理复杂的内窥效果。Large - scale volume data such as smoke in the virtual scene need to be visualized in real - time in the simulation of disaster. However, existing algorithm cannot process the occlusion and clipping of volume in real -time in the hybrid rendering of both polygon and volume data, and can cause ＂the wrong clip of fragment in the endoscope rendering. A novel GPU based hybrid rendering algorithm is proposed in the paper, which treats the volume as participate media, and the result of polygon rendering is used as the restriction of volume integral region. We generalize the clipping to volume rendering and integrate the hybrid rendering seamlessly. Experimental result demonstrates that the algorithm can render the large volume data in the virtual scene and endoscope in real -time.

关 键 词：体绘制 混合绘制 光线投射 体裁剪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]