Research on high resolution remote sensing image storage and efficient management technology in cloud computing environment

Research on high resolution remote sensing image storage and efficient management technology in cloud computing environment

Zhejiang University Kang Jun

The main research contents of this paper are as follows: (1) design of high resolution remote sensing image storage model in cloud computing environment C-RSM based on the analysis and comparison of the current main cloud platform, this paper puts forward the integration of the existing cloud platform Hadoop and eucalyptus, and around the remote sensing image data sharing and map services. , the paper designs the high resolution remote sensing image data organization method based on the Hadoop cloud platform, the version change management mechanism, and proposes the high resolution remote sensing image data partitioning and storage strategy under the Hadoop cloud platform, and designs the high resolution remote sensing image access algorithm under the Hadoop cloud Platform, The access algorithm mainly discusses two algorithms, namely, the high resolution remote sensing image access algorithm under Hadoop, and the access algorithm of remote sensing image Map service under Hadoop is designed based on the Pyramid Model index based on block in Hadoop. (2) based on the C-RSM model, a high resolution remote sensing image management platform is designed C-RSMP based on the analysis and comparison of the spatial data management technology in the current distributed environment, the advantage of using cloud computing technology to manage high-resolution remote sensing images is proposed, and after analyzing the current development of cloud GIS, Proceed to design C-RSMP architecture, service structure, and high resolution remote sensing Image basic service in C-RSMP, including high resolution remote sensing image data sharing service, map service, and high performance computing service in cloud computing environment. In the service design, the paper first proposes to combine the parallel computing ability of GPU with cloud computing technology, and based on the design of GPU accelerated image resampling algorithm and pyramid model creation algorithm, the GPU accelerated image resampling algorithm and high resolution remote sensing image cache management algorithm are designed. In addition, the research contents of high-performance computing services are divided into distributed tasks based on Eucalyptus virtual resource management and High-performance computing based on MapReduce. (3) Combining C-RSMP with land use planning business to realize prototype experiment system: This paper discusses the deployment of system architecture and cloud computing environment in the system, realizes and shows the function modules of land use planning, cloud Platform management, high resolution remote sensing image data sharing, map service and high Performance computing service. The paper chooses some land use planning data of Zhejiang Province to carry out efficiency test, mainly aiming at the high resolution remote sensing Image Access service, map service, and the key algorithm of high performance calculation in this experimental platform, the experiment shows that the algorithm in the platform is correct and the performance is reliable. The research results show that the research method of high resolution remote sensing image storage model and management platform based on remote sensing image application in cloud computing environment can solve the problem of remote sensing image distributed storage and management. Two kinds of high-performance computing services are proposed in this study, in which High-performance computing based on virtual resource management can provide the basis for other serial algorithms to be ported to this platform, and based on MapReduce GPU PlusThe fast remote sensing image algorithm can provide reference for other remote sensing image processing serial algorithm porting to this platform. This paper combines C-RSMP with land use planning industry, which can also provide a new way for other industries to migrate to cloud computing environment.

Keywords: cloud computing High resolution remote sensing image land use planning storage and management High Performance computing virtualization MapReduce Hadoop Eucalyptus

[Download Address]:http://bbs.chinacloud.cn/showtopic-13310.aspx