基于GPU的遥感图像融合并行算法研究

Study of remote sensing image fusion parallel algorithms based on GPU

基于通用GPU并行计算技术,结合遥感图像数据融合处理特点,利用NVIDIA公司的CUDA编程框架,在其GPU平台上对BROVEY变换和YIQ变换融合算法进行了并行研究与实现。实验结果表明,随着遥感图像融合算法的计算复杂度、融合处理的问题规模逐渐增加,GPU并行处理的加速性能优势也逐渐增大,GPU通用计算技术在遥感信息处理领域具有广阔的应用前景。

Based on the GPU parallel computing technology, this article combined remote sensing image data fusion processing characteristics using NVIDIA＇s CUDA （Compute Unified Device Architecture） programming framework with BROVEY and YIQ fusion algorithm. Experimental results show that the computational complexity of the remote sensing image fusion algorithm and the size of the problem increase, the accelerated performance advantage of GPU increases. This study shows that the computing technology of GPU has broad application prospect in the field of remote sensing image processing.