Region growing based improved SAR speckle filter for polarimetric data

Complete polarimetric information allows for better target characterization. However, the polarimetric SAR system suffers from an increase in complexity. Dual-pol imaging modes operate at lower pulse repetition frequencies and hence have a higher swath coverage as compared to quad-pol imaging modes. Hybrid polarimetric techniques have been developed to address the limitations of quad-pol SAR, while still providing comparable target characterization. Speckle is an inherent part of SAR data which causes difficulties in image interpretation and for the extraction of meaningful information. In this paper, we propose an improvement to Intensity Driven Adaptive Neighbourhood (IDAN) speckle filter. Point target preservation technique is also adapted. This filter is analysed using RISAT-1 hybrid polarimetric data acquired over Mumbai, India. The filtered results using proposed filter revels that the dark areas are filtered and strong reflective scatterers are preserved.