A fast normalized cross-correlation algorithm for InSAR image subpixel registration

A fast normalized cross-correlation algorithm (Fast-NCCA) is proposed for InSAR image subpixel registration considering the interferometric cases of gentle topography and/or short baseline. The conventional normalized cross-correlation algorithm (NCCA) searches for the optimal subpixel offset by oversampling either the coarse cross-correlation or the InSAR image pair. The offset estimation accuracy is restricted by the oversampling rate and the computational burden is heavy when high accuracy is demanded. In this paper, the oversampling and correlation searching process of NCCA is transformed into a nonlinear optimization problem which takes the coherent cross-correlation as the objective function, and by solving it with the sequential quadratic programming algorithm, the subpixel offset between the images can be fast and accurately obtained. Experiment on SIR-C/X-SAR data shows that the proposed algorithm performs very well not only on the speed but also on the accuracy.