A NEW INSAR COREGISTRATION STRATEGY FOR GEOPHYSICAL APPLICATIONS

SAR image coregistration is a key procedure before starting InSAR time-series analysis. However, many geophysical datasets suffer from severe decorrelation problems due to a variety of reasons, making precise coregistration a difficult task. Here we present a new strategy that combines pixeloffset series on detected point-like match patches and geometrical mapping functions, in an attempt to improve SAR image coregistration and InSAR time-series analysis. First, all potentially coherent image pairs are coregistered. Then pixel-offset series on point-like targets, detected using coherence analysis, are estimated from amplitude maps and they are used to correct the geometrical mapping function. Finally, all the slave images are resampled depending on the pixel offsets calculated from the corrected geometrical mapping function. The proposed method was applied to two ENVISAT ASAR test datasets and a clear coregistration improvement was found, which should lead to better InSAR time-series analysis results.