SAR Image Classification Using Markov Random Fields with Deep Learning

Classification algorithms integrated with convolutional neural networks (CNN) display high accuracies in synthetic aperture radar (SAR) image classification. However, their consideration of spatial information is not comprehensive and effective, which causes poor performance edges complex regions. This paper proposes a Markov random field (MRF)-based algorithm for SAR classification fully considers the constraints between superpixel Firstly, initialization region labels obtained by CNN. Secondly, probability constructed to improve distribution relationships adjacent superpixels. Thirdly, novel region-level MRF employed classify superpixels, combines intensity one framework. In our algorithm, generation superpixels reduces misclassification at pixel level, rectified improvement description. Experimental results on simulated real images confirm efficacy proposed