Multi Frequency Polarimetric Decomposition of UAVSAR Data

Abstract

UAVSAR is an airborne SAR System which operates in various frequencies of microwave. The UAVSAR is developed by NASA/JPL. UAVSAR data has quad pol capability. Quad pol SAR data or fully polarimetric SAR data (having polarization channels as HH, HV, VH and VV) has capability of distinguishing the geographical features. Different geographical features behave differently for different wavelengths such as for P Band, L Band, S Band, etc. This study is focused on the differentiation of the scattering behavior of different object under different wavelengths. In this research PolSAR Data have been used to characterize the scattering behavior of the objects. It has been found that model-based decomposition techniques provide result of different scatterers according to the mathematical model used in the approach. This causes variation in the scattering values for the same features using decomposition modelling. To overcome this problem, roll invariant parameters have shown their potential over decomposition model to get unique scattering characteristics of the targets. In this research we have worked upon the data of P band and L band dataset of the UAVSAR for the same geographical location located near Candle Lake in Canada. The feature like water body is clearly visible in both the dataset while the vegetation i.e. forest patch is clearly visible in the L band due to the less penetration of the L Band EM wave compared to P Band and the sub-canopy features are better distinguishable in the P Band dataset because of its ability to penetrate through the canopy of the tree.