Potential performance of polarimetric reference function of SAR data processing by coherent target decomposition

Abstract

Significant employment of polarimetric synthetic aperture radar (SAR) is the classification of the Earth’s surface, which includes several processing algorithms, labelling, and representation options. This paper investigated the potential performance of the polarimetric reference function in SAR data processing to apply in coherent decomposition. The proposed methodology employed the simulated radar cross section of dihedral and trihedral corner reflector under a SAR geometry using physical optics (PO) and geometrical optics (GO) to produce the polarimetric reference function for azimuth compression in the processing of SAR data. The Pauli and the Krogager decompositions processed the compressed SAR data to investigate the effect and performance of the polarimetric reference function. The colour clustering results based on k-mean clustering algorithm could deliver more related scattering characteristic. The image pixel classification was strongly associated with the polarimetric reference function in processing step. These findings indicated that the obtained decomposition and clustering results could provide a better performance in visual impact and quantitative evaluation according to the decomposition term, as well as its physical interpretation.