Research On Remote Sensing Image Storage System Based On NoSQL

With the rapid development of aerospace technology and sensor technology,China has launched a variety of high-resolution satellites.These remote sensing satellites have played a major role in resource exploration,disaster prevention and control,and military observation.At the same time,how to effectively store remote sensing images produced by remote sensing satellites has the following challenges: First,the single remote sensing image is large in volume,and even reaches hundreds of GB in some scenarios.Second,the overall scale of remote sensing image data is massive.Third,the real-time requirements of remote sensing image applications are high.According to the above problems,this paper focuses on the image partitioning,image coding and image data access efficiency in distributed storage,and researches the remote sensing image storage system based on NoSQL database HBase,and gives a detailed design scheme.The main results of this paper are as follows:(1)Research on existing remote sensing image storage solutions at home and abroad,and analyze some of the problems.Based on the research of image pyramid technology and GeoHash coding technology,a scheme of image sampling and segmentation coding is proposed.(2)In-depth study of the HBase storage model,based on GeoHash block-encoding proximity and HBase byte-sorted storage features,proposed Rowkey splicing scheme and secondary index based on metadata such as sensor type,after experimental testing,The local index performance of image data has been significantly improved.(3)The reading speeds of MEM,SSD and HDD storage media are analyzed.Based on the LFU data block elimination algorithm,a three-level cache queue is designed.The test data shows that the scheme can effectively improve the access speed of image blocks.And the research and analysis of the Bloom filter principle,designed the Bloom filter module,can quickly detect the existence of image data blocks,avoiding the cost of system resources to find invalid data blocks.(4)The detailed design scheme of the image storage system is given.At the same time,the single point failure problem of HDFS module,HBase database and third-level cache module is analyzed.The construction process of Zookeeper high availability system based on distributed coordination system is detailed.