An autofocus approach for residual motion errors with application to airborne repeat-pass SAR interferometry

Airborne repeat-pass SAR data are very sensible to sub-wavelength deviations from the reference track. To enable repeat-pass interferometry a high-precision navigation system is needed. Due to the limit of accuracy of such systems, deviations in the order of centimeters remain between the nominal and the processed reference track causing mainly undesirable phase undulations and misregistration in the interferograms, referred as residual motion errors. Up to now only interferometric approaches, as multi-squint, are used to estimate those deviations to compensate for such residuals. In this paper we present for the first time the use of the Autofocus technique for residual motion errors. A very robust autofocus technique has to be used since the accuracy of the estimated motion has to be at millimeter scale. Because we deal with low-altitude-strip- map mode data we propose a new robust autofocus technique based on the WLS (Weighted Least-Squares) phase estimation and Phase Curvature Autofocus (PCA) extended to the range- dependent case. We call this new technique WPCA (Weighted PCA). While the multi-squint approach is only able to estimate the baseline variation from coregistered images, the autofocus approach has the advantage of being able to estimate motion deviations independently for each image. Repeat-pass data of the E-SAR system of the German Aerospace Center (DLR) are used to demonstrate the performance of the proposed approach.