摘要： 提出一种针对计算统一设备架构(CUDA)存储器的访存优化策略。在此基础上，给出适用于CUDA存储器的改进空体素跳跃法，以减少对空体素采样点的处理，加快基于Phong光照模型的光线投射算法的运行速度。实验结果表明，该算法能在保证图像质量的前提 下，提高图像绘制速度。

关键词: 光线投射, 计算统一设备架构, 光照模型, 空体素跳跃

Abstract: This paper designs a Compute Unified Device Architecture(CUDA) storage models optimal strategy is designed, and also an improved empty voxel skipping method applied to CUDA storage model is presented to reduce the operation for skipping the empty voxel. It realizes high quality ray casting accelerated algorithm based on Phong lighting model. Experimental results illustrate that the algorithm can significantly accelerate the render speed of ray casting and maintenance the high quality of render image simultaneously.

Key words: ray casting, Compute Unified Device Architecture(CUDA), illumination model, empty voxel skipping

中图分类号: 

• TP391.4