An Improved Processing Method for Low-Frequency Spaceborne Full-Polarimetric SAR Data Affected by Ionospheric Faraday Rotation

Spaceborne full-polarimetric (FP) synthetic aperture radar (SAR) system operating at low frequency (such as L- or P-band) is a powerful microwave sensor used to capture the Earth’s surface information. However, the application effectiveness of the data acquired by such a sensor depends on the processing for the effects of ionospheric Faraday rotation (FR) and possibly-existing certain system errors. This article develops an improved processing method for low-frequency spaceborne FP SAR data affected by FR in the presence of additive noise to generate the 3 × 3 coherency matrix (CM) used for polarimetric analysis. The key to this method is an improved dichotomy of the FR corrected 4 × 4 CM obtained by directly correcting the measured 4 × 4 CM with the estimated FR. The dichotomy is obtained by solving an optimal dichotomy of the FR corrected 4 × 4 CM based on the Schur complement of a 4 × 4 matrix under two constraints with clear physical significance. Specifically, the two decomposed components are Hermitian and positive semidefinite, while the component related to additive noise possesses the smallest trace. From the perspective of polarimetric scattering entropy and the trace of the second component, the effectiveness and robustness of the improved processing method featuring the proposed dichotomy is tested on diverse L-band ALOS PALSAR and ALOS-2 PALSAR-2 FP acquisitions by comparing it with the processing method holding the same basic steps but applying the commonly used direct dichotomy. Experimental results indicate that the improved processing method can effectively alleviate the influence of additive noise on the generated 3 × 3 CM.