High-Resolution Image Data Fusion System Design And Implementation Based On Distributed Parallel Computing

Since the United States launched the first Earth Resources Technology Satellite in the late 1960s,remote sensing entered a new era. With the rapid development of remote sensing technology,it is becoming more and more convenient to get the remote sensing images,which makes the amount of remote sensing data acquisition to show explosive growth,forming GB,TB or even PB-level data.How to effectively store and manage these data and realize rapid sharing and distribution has become one of the focus concern of the spatial information science and some departments.In 2006,the State Council promulgated the "National Long-term Scientific and Technological Development Plan(2006-2020)".In the plan.it clearly establishes the High-Resolution Earth Observation System as one of sixteen major projects.The construction of High-Resolution Earth Observation System has two main objectives:First,build a high-resolution formation of advanced observation systems around the clock,all-time,global coverage of the Earth Observation capabilities;Second,integrate and improve ground resources,establish a center of data and application,and form the chain of space information industry.The goal of High-Resolution Earth Observation System shows that not only storage of high-resolution remote sensing image is a concern,and image processing deserves more attention.According to the processing of the high-resolution No.1, pretreatment is a indispensable part.In this part.on account of the complexity of image fusion algorithm and large time-consuming,the single processing requires a lot of manpower and material resources,So finding a reasonable solution to solve the problem of image fusion is very meaningful. High-Resolution Image Data Fusion Based on Distributed Parallel Computing is to put forward to solve this problem.The technical framework used in this paper is JAVA+ Hadoop+ ENVI, This paper has mainly completed the following researching tasks:firstly, study Several kinds of common remote sensing image fusion which are part of pixel-level fusion in ENVI and draw a conclusion which one is more suitable as a high-resolution remote sensing image fusion method by the experimental result;Secondly, study how to storage and management the high-resolution remote sensing image data based on the HDFS and PostgreSQL;Lastly,study how to implement distributed parallel high image data fusion based on Hadoop and ENVI.