An adaptive polarization calibration method independent of corner reflectors

Abstract

Polarization calibration (PolCal) is necessary for the quantitative application of polarimetric SAR data. The classic PolCal methods rely on corner reflectors (CRs) to calculate parameters, but the deployment of CRs is extremely expensive and they cannot even be deployed in complex terrains. Therefore, currently advanced methods can achieve calibration without relying on CRs. This method is based on the Bragg-like targets and uses the unitary zero helix (UZH) constraint to estimate the co-pol channel imbalance k for the segmented image by Gaussian Newton method. In this process, it is necessary to select appropriate range of k samples for fitting to obtain the final estimate of k. A fixed threshold may lead to improper sample selection, thereby reducing the accuracy of calibration. Therefore, this paper proposes an adaptive PolCal method, which introduces a coefficient of variation to adaptively select samples, and then ultimately estimates k using the best fit line. This paper conducts PolCal experiments using the image of GF3 02 satellite in the calibration field of Etuoke Banner, Inner Mongolia on January 1, 2022. The experiments demonstrate that the proposed method can enhance the stability of the PolCal method independent of CRs and further improve calibration accuracy.