An Improved Decomposition as a Trade-Off Between Utilizing Unitary Matrix Rotations and New Scattering Models

The requirement of an efficient decomposition technique has always been a priority in the field of physical model-based scattering power decomposition. In view of this, the presented work has been carried out in two sections. In the first part, two parallel model-based approaches have critically been analyzed based on their decomposition strategy, result, and limitations. Considering the challenges, the second section emerges out with an improved seven-component ( 7SR) decomposition technique. The proposed architecture is ideally based on embedding the best attributes of both the parallel approaches under one roof. The productivity of the introduced method has been validated on ALOS PALSAR-2 datasets over the San Francisco Bay Area, CA, USA, and Mumbai, India. Algorithmic simplicity and proper distribution of decomposed scattering powers are maintained as the results have been obtained without considering any power restriction constraints. The work shows an enhancement in the underlined dominant scattering mechanism. Simultaneously, a tremendous suppression in the percentage of negative scattering pixels is also achieved, making the 7SR algorithm special among its class.