SPARK Processing of Computing-Intensive Classification of Remote Sensing Images: The Case on K-Means Clustering Algorithm

High performance processing of remote sensing images is an important topic in remote sensing applications. One typical type of remote sensing processing is the iterative computing algorithms such as image classification algorithms, which are often computing-intensive and time-consuming. Recent advancement of cloud computing technologies such as APACHE SPARK has shown great promise for improving the computing performance. This paper presents a MapReduce based approach for parallelizing classification algorithms of remote sensing images on the cloud computing platform. The iterative processing is transformed into iterative Map and Reduce tasks that can be executed in parallel. The K-Means clustering algorithm is experimented with the SPARK cluster deployed on the OpenStack cloud computing platform to illustrate the applicability and effectiveness of the approach.