摘要
为了在大规模场景中渲染出高效率和高真实感的全局光照效果,以图形处理器(GPU)渲染管线流程为可编程基础,利用球面和谐函数进行解码,并且在预计算过程中通过使用小波重建高频部分的信号,来解决预计算辐射传递(PRT)计算过程中丢失掉的高频信号,以免丢失细节变化。在实时渲染全局光照过程中,利用大规模场景的可见性信息进行自适应的细分,使得绘制效率提高。实验结果表明该方法在仿真系统下能够高效真实地渲染出全局光照效果,有较高的绘制效率和真实感。
This paper presented a method of global illumination in Graphic Processing Unit(GPU),which was based on the GPU-programmable rendering pipelines,combined with Pre-computed Radiance Transfer(PRT) formula and decoded by spherical harmonic function,and then the real-time global illumination was rendered.In the process of the pre-computation,the high frequency signal was reconstructed by wavelet,so as not to lose the high frequency signal and the simulation scene was of high reality.According to the information of the visibility for the large-scale scene,the large-scale scene was subdivided by using the method of adaptive subdivision,and the rendering has high efficiency.The experimental results show that the proposed method can generate the global illumination quickly in the simulation system and it has high efficiency and better realistic quality.
出处
《计算机应用》
CSCD
北大核心
2010年第A12期3298-3300,共3页
journal of Computer Applications
基金
国家863计划项目(2009AA01Z332)
关键词
图形处理器
全局光照明
预计算辐射传递
小波处理
自适应细分
Graphic Processing Unit(GPU)
global illumination
Pre-computed Radiance Transfer(PRT)
wavelet processing
adaptive subdivision
