基于GPU加速的光线跟踪渲染算法研究  

Research on Ray Tracing Rendering Algorithm Based on GPU Accelerated

作　　者：陈昱[1] 江兰帆[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350108

出　　处：《武夷学院学报》2013年第2期83-88,共6页Journal of Wuyi University

基　　金：福建省教育厅科技项目(项目编号:JB09017)

摘　　要：分析了光线跟踪渲染算法在GPU上实现的流程;设计了CUDA平台上基于BVH加速结构的光线跟踪渲染引擎,针对CUDA架构的特点重新设计了BVH遍历算法和间接反射计算算法;并在四个场景和两种分辨率下针对GPU和CPU实现进行了详细对比测试。通过分析实验结果,基于GPU加速的光线跟踪算法在不同复杂度的场景下相比其CPU实现都有较高的性能提升。This paper analyzes the process of ray tracing rendering algorithm based on GPU implementation. It designs the CUDA-based rendering engine using BVH acceleration structure. According to the features of CUDA architecture, it redesigns the BVH traversing and indirect reflection calculation algorithm. Finally, it tests the implementations on GPU and CPU under four scenes and two kinds of resolution. Through the analysis of experimental results, ray tracing algorithm based on GPU has significant performance improvement in different complexity scene.

关 键 词：GPU CUDA 光线跟踪 BVH 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]