High Resolution Remote Sensing Image Processing Research Based On Hadoop

Since the 1960 s since its first satellite launch,human exploration of space technology has not stagnation,earth observation technology has also got rapid development,by the hundreds of meters to now the ground of the original resolution has reached 0.61 meters.Due to the high resolution satellite remote sensing data object recognition ability,wide coverage,high data acquisition cycle short,access to the advantages of the diversity,in agriculture,forestry and detection,weather forecast,disaster prevention and mitigation,environmental monitoring,etc,play an important role.With the continuous development of high resolution remote sensing,observation method,to obtain high resolution remote sensing data will be more and more,the application scope will be more and more widely.But in front of the multiply data and user groups of professional development,how to process the data,accurate,effective and timely has become one of the bottleneck of the application of high resolution remote sensing data.Traditional approach for time consuming have can't meet the demand of users,urgent need to improve this dilemma.According to the above problem,based on the analysis of the present situation of high resolution remote sensing data processing on the basis of concrete was carried out by the following several aspects:First of all,according to the present situation of high resolution remote sensing image processing and the cloud computing technology into the data processing,has set up a platform based on Hadoop of high resolution remote sensing image processing architecture,and the framework are described.In view of the characteristics of high resolution remote sensing images at the same time,put forward suitable for high resolution remote sensing image processing on Hadoop platform shard strategy execution.Second,choose the Meanshift algorithm to build the high resolution remote sensing image processing based on Hadoop platform architecture,on the basis of the Meanshift algorithm principle,a mathematical formula in Hadoop platform implementation.Finally,the experiment platform of building large data processing test and Meanshift algorithm for image segmentation experiments,and from the results and performance analysis of two aspects of the experimental results.The experimental results show that the built hadoop framework can undertake large data processing,and processed images meet the definition of remote sensing image segmentation,Hadoop environment for high resolution remote sensing image processing,can effectively reduce the overall processing time.