基于GPU的图像处理并行算法分析

Analysis of Parallel Image Processing Algorithm Based on the GPU

在计算机技术的快速发展过程中,人们需要对各种复杂的图像进行处理,图像处理的方法呈现出多样化和复杂化的特点,人们对CPU的性能要求也在逐渐提高,传统的CPU图像处理方法已经不能满足目前人们的实际需求。对一种高效快捷的图像处理方法进行研究,在充分利用GPU并行处理能力的基础上,采用G++语言来完成相应的图像计算,其中的图像处理并行算法主要包括彩色负片处理算法、透明合并处理算法等,并且与GPU实现相同效果的性能进行相互对比,以此来证明基于GPU的图像处理并行算法的高效性。

In the process of computer technology rapidly develop, people need to deal with all kinds of complex images. Image processing method presents diversity and complex characteristics, and the remind for CPU capability is higher, the traditional CPU image processing already can not satisfy the needs of people at present. In this paper, it research a high effciency and fast image processing method. It makes full use of OPU parallel processing ability, using G＋＋ language to complete the corresponding images. The image parallel processing algorithm method main contain the color negative film processing algorithms, transparent merging processing algorithms and so on, and comparing the performance of achieving the same effect with the GPU, to prove the efficiency of parallel image processing algorithm, based on GPU.