Research On Management And Service Technology Of Remote Sensing Image Big Data Based On GeoTrellis

With the development of remote sensing Earth observation technology,multidimensional,multi-scale and multi-temporal remote sensing image data are continuously generated.The data volume of remote sensing image data has increased from the early GB level to the current TB,PB and even ZB level.Efficient organization of massive remote sensing image data has gradually attracted the researchers’ attention.At present,the traditional remote sensing image data management model can not meet the requirements of remote sensing image big data organization and management.Combining emerging data technology and management mode to develop management organization programs for remote sensing image big data has become a hot research spot in academia.Based on the existing remote sensing image data organization and management technology,this paper proposes an efficient,stable and reliable remote sensing data management service solution based on distributed storage and computing technology providing a new way to solve the organization and management of massive remote sensing image data.The new solution base on the distributed file system HDFS and the GeoTrellis which is developed on Spark.And the solution also uses tools such as Accumulo and YARN and so on.The purpose of the solution is to achieve stable and efficient organization and management of remote sensing image big data.In the process of image data management,the hybrid storage model of remote sensing data and the spatial and temporal indexing technology of remote sensing image tiles are introduced.In the process of service construction,a pattern conversion program is designed for the problem of tile service format.In this paper,a remote sensing data management service prototype system is realized by combining the design scheme with the actual remote sensing image big data management’s requirements.The prototype system solves the problems encountered in remote sensing image big data management services from three main aspects: remote sensing image storage and downloading,image query and image service publishing.Through the performance test of the system,the remote sensing image pyramid model of the prototype system is more than twice as efficient as the ArcGIS Server system,and the stability of the prototype system pyramid model is verified by increasing the size of the constructed image.In addition,the stress test proved that the prototype system remote sensing image tile service has the ability to process more than 100 transactions per second when the simultaneous access user is 100.The test experiments verify the feasibility of the GeoTrellis-based remote sensing big data image management service program.In general,remote sensing big data image management service solution based on GeoTrellis meets the main needs of remote sensing image big data management services and has obvious advantages over traditional solutions.The high efficiency,stability and reliability of the prototype system verify the practical value of the system scheme and provide strong support for the promotion and application of the system scheme.