大规模测量平差分布式计算技术研究

项目名称： 大规模测量平差分布式计算技术研究

项目编号： No.41274015

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 吕志平

作者单位： 中国人民解放军信息工程大学

项目金额： 70万元

中文摘要： 现代大地测量数据处理具有规模大、数据多源多期、精确性与实时性要求高等特点，缓解海量数据处理与有限的计算、存储能力的矛盾，实现高性能分布式计算，是现代大地测量重要的研究方向之一。目前，以网格计算和云计算为代表的新型分布式计算技术以其高效的资源利用和计算效率而逐渐成为解决大规模数据计算的首选方法。本项目针对建立大规模测量平差分布式计算系统这一目标而展开，研究基线解算的分布式计算方法、法方程层面和观测方程层面的测量平差分布式计算模型、大规模测量平差分布式计算的质量控制方法，开展测量平差分布式计算原理性程序设计和试验。突破测量平差的并行计算算法、分布式平差节点质量控制方法等关键科学问题，实现大规模大地测量网络多节点协同计算方法的创新。本项目将为提高大地测量数据处理效率，建立一体化、自动化、智能化的大地测量服务系统提供应用基础。

中文关键词： 分布式计算；基线解算；测量平差；计算模型；质量控制

英文摘要： The data processing of modern geodetic survey bears characteristics of large scale, multi-source and multi-period data, high accuracy and instantaneity and so forth. Thus it has become an important research field to alleviate the contradiction between the massive data processing and the limited computing and storage capacity so as to achieve distributed computing with high performance. At present the new distributed computing which is represented by grid computing and cloud computing is turning into the preferred means for solving large-scale data computing with its efficient use on resource and computational efficiency. The project aims at establishing a distributed computing system for large-scale survey adjustment to carry out related researches, including distributed processing methods of the baseline solution, distributed computing models for survey adjustment based on the layer of the normal equation and observation equation, quality control methods for distributed computing of large-scale survey adjustment. A distributed computing original rational program of survey adjustment is finally developed and tested. It solves some key scientific problems such as parallel computing algorithms of survey adjustment and quality control methods of distributed adjusted nodes to realize methods innovation on multinodes

英文关键词： distributed computing；baseline solution；survey adjustment；computing model；quality control