Applying deep learning for agricultural classification using multitemporal SAR Sentinel-1 for Camargue, France

摘要: The aim of this paper is to provide a better understanding potentialities the new Sentinel-1 radar images for mapping different crops in Camargue region South France. originality relies on deep learning techniques. analysis carried out multitemporal data over an area Camargue,France.50 processed order produce intensity stack from May 2017 September 2017. We revealed that even with classical machine approaches (K nearest neighbors, random forest, and support vector machine), good performance classification could be achieved F-measure/Accuracy greater than 86 % Kappa coefficient 0.82. found results two recurrent neural network (RNN)-based classifiers clearly outperformed approaches. Finally, our analyses show same was obtained RNN-based Rice class, which most dominant crop region, F-measure metric 96 %. These thus highlight near future, these techniques will play important role remote sensing time series.