Some concepts for precise estimation of deformations/rigid areas in polar pack ice based on time series of ERS-1 SAR images

Abstract

This paper describes concepts for an algorithm using pairs of satellite SAR imagery to estimate precisely rigid areas and deformations in polar ice fields. The idea is first to produce a sparse sampling of the ice displacement and then identify rigid subsets of this sparse sampling. The algorithm tunes the position parameters of the images using those of these subsets best fitting the model of rigid motion (shift/rotation). Then the algorithm identifies rigid areas again and so on iteratively improving the identification of rigid areas and the relative geographical positioning in the images. The last step is to classify each pixel in the images as included in one of the given rigid areas or, perhaps, not included in one of these areas.