Research On Massive Image Tile Storage Method Based On Object Storage Structure

As one of the important data sources in scientific research and industrial application,remote sensing image data has the advantages of global coverage,rich information and fast acquisition,which has been widely applied to disaster monitoring,meteorological forecasting,resource survey,environmental protection and many other fields.Remote sensing tile data,as one of the main service methods of remote sensing image,has become more and more popular in the sharing,query and browsing of remote sensing image data.In recent years,with the rapid development of high-resolution earth observation technology,the amount of remote sensing image data has increased geometrically,which has also brought about an explosive growth of image tile data,bringing storage for image tile data huge challenge.Therefore,it is of great significance to implement the storage organization of tiles and realize the rapid response and acquisition of massive tile data.On the basis of the existing tile storage methods,this paper designs a massive tile data storage method based on object storage structure,aiming at the shortcomings of metadata management,system resource consumption and request response speed in tile data storage.The main research contents are as follows:(1)Massive tile data storage framework design based on object storage structure.In order to meet the storage requirements of massive image tile data,based on the in-depth study of image tile data features and object storage structure characteristics,this paper introduces the object storage idea into the tile storage domain,and fully considers the characteristics of tile data,designing the overall structure and read and write flow of massive tile data storage.(2)Research on inter-node data distribution and intra-node organization strategy for spatial tile objects.In order to realize the efficient reading efficiency of spatial tile data in the object storage cluster,this paper considers the distribution of tiles between multiple object storage devices and the organization within a single object storage device,designing two-wheel mapping tile distribution strategy based on aggregation interval and tile organization based on compact storage.(3)Research on tile caching strategy based on theme temporal and spatial value.In order to improve the request response speed of tile data,based on the research of tile data access characteristics and existing cache strategy,this paper designs a two-level cache index based on topic pyramid and a cache replacement process based on theme temporal and spatial value.(4)Prototype system design and experimental verification.Based on the above model framework and key technologies,this paper designs and implements a prototype system based on the framework,and uses the prototype system to carry out experimental verification of key technologies.The results show that the tile organization strategy based on compact storage and the two-wheel mapping distribution strategy based on aggregation interval can effectively improve the tile response speed,while the tile cache strategy based on theme temporal and spatial value has the higher cache hit ratio than the traditional cache strategy.