Multi-Class Change Detection Using Polarimetric SAR Data

Remote sensing, particularly Synthetic Aperture Radar (SAR) systems, has lots of operational advantages in continuous change monitoring with its all-weather, day and night imaging capability. In particular, the fully polarimetric SAR systems can offer a more efficient and reliable means of change monitoring by providing additional observations about the Earth's surface. Recently, previous change detection techniques using SAR data focus on detecting changed areas from non-changed areas. However, the utilization of microwave scattering information can offer complementary information on various scatterers and can be a potential capability in discriminating change types without in-situ information. Therefore, this study aims to propose a new multi-class change detection method based on the fusion of polarimetric features. The multi-class change detection results show that the proposed method can discriminate the various land cover type changes in terms of radar scattering mechanism.