Big data streaming for remote sensing time series analytics using MapReduce

Governmental agencies provide a large and open set of satellite imagery which can be used to track changes in geographic features over time. The current available analysis methods are complex and they are very demanding in terms of computing capabilities. Hence, scientist cannot reproduce analytic results because of lack of computing infrastructure. Therefore, we propose a combination of streaming and map-reduce for time series analysis of time series data. We tested our proposal by applying the classification algorithm BFAST to MODIS imagery. Then, we evaluated account computing performance and requirements quality attributes. Our results revealed that the combination between Hadoop and R can handle complex analysis of remote sensing time series.