Method For Storing And Retrieving Massive Multisource Heterogeneous Remote Sensing Image Metadata

Remote sensing image metadata,as a kind of geographic information metadata,describes remote sensing image satellites,sensors,cloud cover,imaging time,geographic range and other information.With the vigorous development of remote sensing technology,the data volume of remote sensing image metadata continues to grow.At the same time,remote sensing image metadata sources are diverse,and different types of metadata files have the characteristics of different structures and different semantics.Remote sensing image metadata has the characteristics of massive,multi-source heterogeneous,written once and read many times.Aiming at the above characteristics,this paper proposes a solution for efficient management of spatiotemporal remote sensing image metadata,and verifies the feasibility of the solution through a prototype system.The specific summary is as follows:(1)In view of the characteristics of remote sensing image metadata,in order to realize the unified storage of remote sensing image metadata,an integrated storage strategy of remote sensing image metadata that takes into account the spatial characteristics,a segmented storage strategy of remote sensing image metadata that takes into account the heterogeneous features of multiple sources,and a partitioned storage strategy of massive remote sensing image metadata are proposed.In addition,a dynamically expanding remote sensing image metadata analysis and extraction method is proposed to quickly support the import of new types of metadata.(2)By extending the function of the generalized search tree and adding a time dimension to the spatial index,a spatiotemporal hybrid index of remote sensing image metadata based on the generalized search tree is proposed to improve the retrieval efficiency of metadata.A series of metadata retrieval experiments were carried out to verify the effectiveness of the remote sensing image metadata organization model and the advantages of spatiotemporal hybrid indexing in retrieval efficiency.Compared with the spatial index,the spatiotemporal hybrid index improves the retrieval efficiency by about 4 times.(3)Constructed a remote sensing image metadata management prototype system using the open-source framework and class library.The prototype system applied the remote sensing image metadata integration,segmentation,and partition storage strategy,and realized the unified organization and storage of remote sensing image metadata.Optimization schemes such as spatiotemporal hybrid indexing are used to achieve efficient retrieval of remote sensing image metadata.