Remote sensing algorithm platform in Windows Azure

Abstract

As a kind of eScience, remote sensing needs numerous computing resources to process data. A lack of computing resources restricts many scientists' work. The advent of cloud computing solves the problem perfectly for its low-cost and highly scalable computing power. This paper introduces the remote sensing algorithm platform running on Windows Azure. Windows Azure provides the relevant algorithm and efficient and extensive computing resources to solve large scale remote sensing image processing computations for myriad researchers. This platform applies the MapReduce model that constructs the parallel data processing module to organize and coordinate work flow among virtual machines. Efficiency tests show that by using the MapReduce model, the remote sensing algorithm platform efficiency in data processing has been dramatically improved. This paper relays the experience of using Windows Azure in eScience scenarios similar to remote sensing for reference in future research.