Research On The Key Techniques Of Multi-source Remote Sensing Big Data Management Under The Cloud Computing Environment

With the rapid development of remote sensing satellite technology,not only caused the sustained growth of satellite and data type number,but also led to the great increasement of remote sensing data volume,which brought greate challenges to the data management,data storage and data organization.Furthermore,for the reason that remote sensing data own the multidimensional characteristics such as space,time and so on,more and more remote sensing applications start focusing on data analysis and computing by the space or time dimension.But under existing data center conditions,it is very inefficient for data acquiring,sharing and analysis.And how to provide users the efficient services of data on-demand subscribing,sharing and space-time computing has been the urgent problems in the field of remote sensing.Therefore,aiming at the problems mentioned above,this paper use the advantages of largescale data storage,virtualization,elastic expansion and on-demand services of cloud computing to study the data storage,organization and acquiring of multi-source remote sensing data under the cloud computing environment,with the support of the National High Technology Research and Development Program of China(“863” Program).This research mainly contains the following points:(1)For the problem of different configurations between multi-source remote sensing data,we introduced the geometric information metadata standard to establish the standard metadata model,and then realized the unfied format conversion and data retrieval of muti-source remote sensing metadata.Moreover,we used different ways to store the remote sensing data because of the difference between archived data and calculated data.(2)For the on-demand service of data acquiring and sharing,this paper first gived the design of virtual directory on the base of object files,and then provided users the independent cloud space which own hierarchical directory,just as file system.We final give the concept of virtual view and virtual mapping,which is used to proviced users efficient services of data subscription and sharing.(3)For the analysis of multidimensional remote sensing data,we combined the technology of Open Data Cube,and provided users the service of multidimensional data cube model construction and time series computing based on memory with the data stored in the cloud storage.Based on the research of the above key techniques,this paper implemented the unified management of multi-source remote data,and we provided the users of remote sensing cloud ondemand services of cloud space management,space-time data subscription,data sharing and time series analysis analysis.And through the relevant functional tests,we validated these services mentioned above.We provided systemic and completed solutions for these problems mentioned above.Innovation of this paper is that we implemented the unified management and retrieval of multi-source remote sensing data based on standard metadata model,and we also realized the ondamand service of data subscription and data sharing face to multiuser based on cloud storage and virtual mapping,which avoid the unnecessary data transmission and duplicate organization.