Multitemporal Polarimetric SAR Change Detection for Crop Monitoring and Crop Type Classification

The interpretation of multidimensional Synthetic Aperture Radar (SAR) data often requires expert knowledge. In fact, it to simultaneously consider several time series polarimetric features understand the physical changes a target and its temporal evolution. To characterise over time, Multitemporal Polarimetric SAR (MTPolSAR) change detection has been introduced in literature [1] [2]. However, previous methods either only exploit intensity or resulting changed scattering mechanisms are not guaranteed represent target. This paper presents variation detector used [2] based on difference covariance matrices that information, allowing for an intuitive representation characterisation dynamics. We show results this method monitoring growth stages rice crops present novel application crop type mapping from MT-PolSAR data. compare performance with neural network-based classifier uses PolSAR derived matrix decomposition as input. Experimental classification proposed baseline comparable, differences between two overall balanced accuracy F1-macro metrics around 2% 3%, respectively. presented here achieves similar performances traditional while providing additional advantages terms interpretability insights about time.