A joint coregistration of rotated multitemporal SAR images based on the cross-cross-correlation

Accurate synthetic aperture radar (SAR) images coregistration is on the base of several remote sensing applications, such as interferometry, change detection, etc. This paper proposes a new algorithm for jointly coregister a stack of multitemporal SAR images exploiting the cross-correlations computed for each couple of patches' cross-correlation. By doing so, the method is capable of exploit also the respective misregistration information between the slave during the estimation process. This methodology is applied to improve the performance of the constrained Least Squares (CLS) optimization method that does not account for the reciprocal information related to the slaves. Tests on real-recorded data shown the benefits of the proposed method in terms of root mean square error (RMSE) for images affected by respective rotations.