Research On High Performance Water Extraction Method Of Remote Sensing Data Based On Cluster System

Due to the development of earth observation, it is easier to get multi-source, high-qualitydata source. Researchers from different countries have paid much attention in extractinginformation from remote sensing image recently. Among these studies, water extraction resultscan be used in lots of areas, such as agriculture, water resources protection and disasterprevention. However, there is a large amount of remote sensing data so far, and it is unable toapply the water extraction method in large-scale. Simultaneously, remote sensing informationcomputation consumes large scale of computing resource and time. It is more difficult to extractwater information because of the uncertainty of remote sensing data. Recently, as parallelcomputing and grid computing become popular and are applied to lots of fields, we can considerthem as a solution to the time-consuming problem in water extraction processing.In this paper, we propose a method which constructs a cluster system and applies the waterextracting algorithm to the cluster environment. As a result, the method is proved to be efficientand can be applied in large-scale application. The main work of this paper is as below:Firstly, a survey of traditional water extracting algorithms is performed and we analyze theadvantages or disadvantages of these algorithms. Additionally, we discuss the procedure andresult of another water extracting method–automatic extract water information through iteration,this method is proved to be more accurate. At last, we analyze several limitation of remotesensing computation and propose a solution–Cluster Computing.Secondly, we propose a method to construct a Cluster system which is suitable for remotesensing information computation. In this section, several strategies such as data storage and dataadjustment are taken into consideration. Additionally, we propose a new water extracting algorithm which runs on a Cluster system and improves the efficiency of water extraction. In thissection, we solve the problem of task partition and load balancing.Finally, we design and implement Environment and Resource Data Management System ofCenter Asia. In this system, it is convenient to extract water in large-scale and proved to beefficient.