Parameters analysis for polarimetric SAR Based on classification accuracy

Abstract

As a multi-channel microwave remote sensing imaging radar system, polarimetric Synthetic Aperture Radar (SAR) enhances the information extraction ability for material scene in the observation area, in which the design and state monitoring for parameters play an important role in polarimetric SAR system and equipments management. This paper presents an analysis method for three typical polarimetric SAR system parameters: channel noise, polarization imbalance, and polarization isolation. We explore the polarimetric SAR image classification accuracy ratio and the relationship with three parameters above, review the statistic model of each parameter, and propose an analysis model between each parameter and the classification accuracy ratio measurement. By the polarimetric SAR image experiment, the comparative curve between each parameter and the influence on classification performance is obtained, which shows an application reference for polarimetric SAR system design and management.