项目名称: 基于剖分模型的遥感影像模板并行处理方法研究
项目编号: No.U1304403
项目类型: 联合基金项目
立项/批准年度: 2014
项目学科: 天文学、地球科学
项目作者: 杜根远
作者单位: 许昌学院
项目金额: 30万元
中文摘要: 遥感影像数据在诸多军事和民用领域有着越来越广泛的应用,随着数据量的增加和应用需求的扩大,遥感数据的组织管理和处理速度已经成为制约遥感技术应用的瓶颈,地球剖分理论和高性能计算为上述问题解决提供了一种可能途径。本项目旨在通过研究剖分遥感影像面片模板的快速并行处理方法来解决这一关键性问题,主要研究内容包括:(1)在地图分幅拓展的地球剖分模型基础上,建立剖分面片模板的数据模型,提出模板快速生成方法,设计剖分面片模板计算模式及模板并行计算机制;(2)在MPI与OpenMP混合并行计算框架下,对模板处理算法进行并行性分析,建立模板快速并行处理模型与方法,设计实现算法的并行化类库;(3)开发剖分面片模板并行处理算法软件平台,利用特定地区遥感影像资源实现模板快速应用模式。本研究对于拓展地球剖分理论应用领域,提高遥感影像处理速度,加快空间信息快速可视化表达、分析及决策反应速度,具有十分重要的战略意义。
中文关键词: 空间剖分;面片模板;并行计算;分层组织;
英文摘要: Remote sensing image data has more and more wide application in many military and civilian areas. With the increasing data amount and the extending application requirements, the organization management and processing speed of remote sensing image data has become a bottleneck of restricting the application of remote sensing technique. Both The Global Partition Theory and the high performance of computing provide a possible solution for the above problem. This project aims to study the fast parallel processing methods of spatial data partition facets to solve the key problem. The main contents include: (1) based on the global spatial data partition model of EMD, propose the fast generation method of templates and design the computing mode of partition facets template and template parallel computing system; (2) on the hybrid parallel compute framework of MPI and OpenMP, analyze the parallelism of template processing algorithms, put forward the parallel method of template processing algorithms, design and realize the parallel class library; (3) develop the parallel processing algorithms software platform of the partition facets template, use remote sensing images of specific areas to confirm the rapid application mode. This research has very important strategic significance for expanding the application field of the
英文关键词: spatial partition;facet template;parallel computing;layer-based organization;