Research On Efficient Management And Scheduling Methods Of Massive Multi-source Heterogeneous Remote Sensing Images

The rapid development of image sensor technology and the increasing number of Earth observation satellites have led to a sharp increase in the amount of remote sensing images.The storage of massive images and the real-time scheduling of gigabyte-level images bring great challenges to computer hardware,network transmission and storage systems.In view of the above problems,this paper carries out research on the efficient management and scheduling method of massive remote sensing images,and the main research contents and achievements are as follows:(1)The advantages and disadvantages of the current mainstream relational database and No SQL database are compared and analyzed,and finally the remote sensing image metadatabase cluster is selected to be built with Postgre SQL.Referring to the metadata standards of remote sensing images at home and abroad,a unified metadata organization structure suitable for multi-source heterogeneous remote sensing images is designed,and the automatic parsing and storage algorithm of metadata is realized.At the same time,a multi-level Hilbert grid coding is designed based on the Hilbert spatial filling curve algorithm,and it is used to construct a one-dimensional spatial index of remote sensing images.(2)The advantages and disadvantages of the current mainstream distributed file system and its applicable scenarios are compared and analyzed,the advantages of HDFS file system as the underlying storage system for massive remote sensing images are summarized,and the shortcomings of HDFS in small file storage are optimized with Hadoop Archive.Based on the five-layer fifteen-level image tile pyramid model,the remote sensing image is sliced and stored,and the image tile segmentation and warehousing workflow is encapsulated to achieve one-click operation.(3)The process of image scheduling is deeply analyzed,divided into three parts:image retrieval,data transmission and image visualization,and the algorithm optimization is carried out for each part.Three image scheduling optimization strategies(ring cache mechanism,tile prefetching mechanism and multi-threaded mechanism)are designed,using the ring cache mechanism as the scheduling data transfer station,using the multi-threading mechanism to realize parallel data reading and writing,and using the tile prefetching mechanism to improve the smoothness of image preview.The three mechanisms work in harmony to greatly improve the performance of image scheduling.(4)Develop a massive remote sensing image management and scheduling system(RSIMSS)based on JAVA development language and Vue front-end development framework.The system adopts a layered modular architecture,which is divided into data layer,service layer,and user layer.The data layer uses Postgre SQL and HDFS distributed clusters to realize the hybrid management of structured data and unstructured data of remote sensing images.The service layer includes an image retrieval module,a data transmission module,and an image visualization module to respond to users’ data requests and complete data scheduling.The user layer hides the complex internal structure of the system,providing users with a series of functions such as data retrieval,download,and preview.Finally,RSIMSS is tested,and the results show that RSIMSS has efficient and stable image storage,retrieval and scheduling performance.Figure [32] Table [8] Reference [92]...