基于机群的并行匹配算法

A parallel image matching algorithm running on cluster

随着卫星遥感技术的发展,需要快速地将卫星遥感图像数据转化为用户需要的信息,并行图像处理技术是解决“快速”的重要途径。并行程序的性能与计算机体系结构密切相关,不但取决于CPU,还与系统架构、指令结构、存储部件的存取速度等因素有关。一般意义上,提高并行程序的性能采用粗粒度并行,指令级优化(ILP)和存储优化等技术。作为尝试,本文讨论了在工业标准化机群上采用软件式共享存储系统做的并行影像匹配方法,以影像匹配算法为例子,讨论了如何在粗粒度并行、指令级优化(ILP)和存储优化三个方面提高图像处理的计算速度。

It is the key issue in remote sensing field to transform massive data into information in short time.Parallel image processing on high performance computing is one of the key technologies to solve this problem quickly.The performance of parallel program is closely related to computer architecture,besides CPU,including system framework,instruction structure and access speed of storage unit.Generally there are several ways to improve the performance of parallel program,such as coarse grain parallelism,instruction level and memory optimizing.In this paper,a parallel image matching algorithm on Cluster is studied.Based on this algorithm,the image processing speed can be improved with the help of coarse grain parallelism,instruction level and memory optimizing.