Research On The RS Partitioning Data Processing Based On The GEOIP Cluster System

With the development of aerospace technology, the data storage of remote sensing increases fast, but on the other hand, the way to store these mass data come to be a critical problem. In this article, I have discussed the mode of storing massive remote sensing data based on global subdivision technology.Generally speaking, the storage mode of image is based on map sheet in the way which no relation lays between any two images. But, in this article, as I subdivide image into pieces by global subdivision method, there has logical relation between each of subdivision pieces, and this is done by the subdivision encoding method. Every image of global subdivision has its own code, the inside of subdivision code has hierarchy of different spatial resolution. And by encoding, this hierarchy relation can pass to the images. In this way, the application of image usage will be fundamentally changed.In this article I build the cluster system based on local distribution method, which has the ability to store massive image data and also has high performance, and its hardware structure is storage area network. In this article, I have done the following research: Use the method of QTM encoding based on the comparison of different subdivision research to encode the subdivision pieces; Put forward the architecture of a subdivision cluster for storing subdivision pieces. And have discussed the configuration of this cluster; Put forward the architecture of a multi-sever metadata server. And have discussed the distribution of metadata in multi-sever based on allocation catalogue technology; Put forward a method to format isomeric remote sensing image. Study the way to format metadata,supplementary information and temporal partition. I also study the transition between latitude and longitude to subdivision encoding; Study the theory of virtual storage, and put forward a local subdivision storage method;...