ABSTRACT
State-of-the-art polarimetric decompositions are functional models that are capable of differentiating polarimetric signatures corresponding to different targets. One of the utilities of polarimetric decomposition is to manifest complex information from SAR images and classify targets/land cover based on the scattering mechanisms they follow, i.e., geophysical characterization of the targets. Incoherent decomposition methods categorize the geophysical attributes of targets by decomposing the second-order covariance/coherency matrix into elementary second-order descriptor matrices. However, polarimetric decompositions have a limitation in terms of target characterization. This chapter will reflect on one of the methods that can be used to curb the limitation of polarimetric decompositions by synthesizing temporal characteristics of scattering phase correlation, i.e., polarimetric interferometric (PolInSAR) coherence with generic decomposition model. This chapter will give a detailed description of the hybrid decomposition model that uses PolInSAR coherence and statistical decomposition parameters (entropy and alpha angle) to characterize natural and man-made targets. The chapter will discuss in detail the experiments of the hybrid model with RADARSAT2 PolInSAR data acquired over the Dehradun region in Uttarakhand, India. The experiment results have been compared with generic and modified decomposition models and will be discussed in this chapter.
