Research On The Pyramid Construction And Storage Technology Of Image Data Based On HDFS

Nowadays the remote sensing image data in the forest resources management and prediction plays more and more important role, and in the face of the data becomes more and more huge, how to effectively manage the remote sensing data is the most need to solve the problem. But in the forestry resource data, not only the remote sensing data, investigation data of forest resources is an important data source in forestry data too, through the integration of resources, comprehensive utilization of data is an important trend in the future. Through distributed technology and domain analysis technology, it is very important to solve the problem of remote sensing data management technology and the construction of the platform.In this paper, we mainly study the parallel construction technology of image Pyramid and the key technology of forestry comprehensive data management. First introduce the quad-tree coding, and through the quad-tree coding for remote sensing data processing. After the coding, through the analysis of the shortcomings of the traditional image Pyramid parallel construction, the strategy of a distributed construction sub Pyramid is designed. On the basis of the MapReduce parallel processing model is used to design the construction of Pyramid, Pyramid finally realize parallel construction.In the aspect of forestry comprehensive data management, the application layer and the bottom structure are analyzed. The application layer explains the construction of the platform through the domain analysis technology, so as to realize the maximization of software reuse. On the underlying structure, propose use Peano space filling curve to index data based on HBase database. By way of the index can guarantee to realize the physical address of the adjacent which is based on quadtree coding data and improve the retrieval efficiency. And the underlying storage structure of the distributed file system is optimized, and the storage space can be effectively utilized. Through the establishment of storage model based on Hbase process the data storage, to achieve rapid storage data under MapReduce, so as to realize the design of integrated management of forestry data.