基于集群的海洋遥感图像融合并行计算策略

集群体系下的大规模并行计算,是高性能计算的基础。遥感图像处理效率的提高,有赖于并行计算技术的应用。在分析已有网格计算环境下分布式任务分配方法的基础上,针对海上遥感图像目标物数量相对较少的特点,首先利用四叉树结构理念对目标区域进行划分,同时采用动态负载均衡的任务分配策略与并行计算思想,提出对目标区域图像进行融合处理的集群体系任务分配算法处理模型。通过对比验证,表明该集群体系下算法模型能有效地提高图像融合的速度。

CLUSTER-BASED OCEAN REMOTE SENSING IMAGE FUSION PARALLEL COMPUTING STRATEGY

Large scale parallel computing under cluster system is the foundation for high performance computing.The enhancement of the efficiency of remote sensing image processing depends on the application of parallel computing technology.This article,on the basis of analyzing distributed task assignment approaches under the existing grid computing environment,aiming at the sparse characteristics of ocean remote sensing image objects,from the software aspect,first of all uses the quadtree structure idea to segment object areas,while at the same time uses the dynamic load balancing task assignment policy and parallel computing thought to present a cluster system task assignment algorithm processing model for object area image fusion processing.By comparative validation the algorithm model under that cluster system is proven to be effective at speeding up image fusion.