基于并行计算的全息图快速生成

Fast Hologram Generation Based on Parallel Computing

全息图的快速生成是当前计算机制全息的重点与难点,为提高全息图的生成速度,提出一种基于Cuda并行计算的全息图快速生成方案。首先利用OpenGL的深度缓存对空间物体进行离散取样,获得空间离散物点集,再对传统的查表算法进行优化处理,大大减小查找表的空间大小,离线制作查找表,存入GPU纹理内存,合理设计并行计算方案,将Cuda并行计算的方法应用于全息图的快速生成。实验表明该方法有效可行,并行计算可将全息图的生成速度提高40倍左右,同时,OpenGL的使用给交互式全息图的计算机生成提供一种研究思路。

Fast hologram generation is the key for the study on Computer Generated Hologram （CGH）. In order to accelerate the generation speed of CGH, a fast scheme based on Cuda parallel computing was proposed. Firstly, the objects were sampled by means of OpenGL depth buffer, and objects points set was acquired. Then with the optimization of traditional look-up table algorithm, storage space was reduced, a short look-up table was produced offline and stored in the GPU texture memory. Finally, a reasonable parallel computing program was designed to generate CGH. The experimental results showed that the algorithm was effective and feasible, and the speed was increased by about forty times with the parallel computing technology. At the same time, the use of OpenGL provided an idea for computer-generated interactive holograms.