A New Unified Approach to Channel Imbalance and Cross-Talk Calibration of Polarimetric Data

A new unified approach to channel imbalance and cross-talk calibration of polarimetric data is described. This approach is applied during the SAR image formation, employing the multilook SAR processing technique and the covariance matrix of observed data. This method makes no assumptions about the reciprocity of the distortion model and noise system, and about the presence of dominating target with azimuthal symmetry. The standard linear and nonreciprocal distortion model for the received signal is assumed for each polarimetric image obtained from three looks without overlapping. This leads to a set of equations that relate the observed covariance matrix and the distortion system parameters. The non-linear solution of the calibration parameters is solved by iterative approach based on numerical methods. Fully polarimetric SAR L-band data, acquired by CENSIPAM Airborne R99-SAR on the Tapajos Forest (Brazilian Amazon), was used to evaluate the method as well the polarization signature response of some trihedral corner reflectors deployed in the area.