Semi-Supervised Land Cover Classification Using Pi-SAR2 Observation Data

The National Institute of Information and Communications Technology (NICT) developed and is operating the airborne X-band polarimetric and cross-track interferometric synthetic aperture radar (SAR): Pi-SAR2. Deep learning is one of the machine learning techniques that has recently been actively used in various fields, including remote sensing. However, it is difficult to prepare sufficient annotated training data for supervised learning. This paper proposes a semi-supervised deep learning method for land cover classification using Pi-SAR2 polarimetric data and reports the results of an experimental investigation.