High Resolution Satellite Image Processing Using Hadoop Framework

Complex image processing algorithms that require higher computational power with large scale inputs can be processed efficiently using the parallel and distributed processing of Hadoop MapReduce Framework. Hadoop MapReduce is a scalable model which is capable of processing petabytes (10¹⁵ order) of data with improved fault tolerance and data parallelism. In this paper we present a MapReduce framework for performing parallel remote sensing satellite data processing using Hadoop and storing the output in HBase. The speedup and performance show that by utilizing Hadoop, we can distribute our workload across different clusters to take advantage of combined processing power on commodity hardware.