Improved Phase Unwrapping Algorithm Based on Standard Methods

This article contains description of an improved tool for phase unwrapping, as developed and to be applied routinely within COMET LiCSAR system producing Sentinel-1 interferograms in 0.001° resolution (WGS-84). We apply existing standard methods on routinely available data in a way to improve reliability of the standard phase unwrapping procedure and increase coverage of unwrapped pixels. The primary idea is to reduce high phase gradients, causing both decorrelation and zonal unwrapping errors, by extracting and unwrapping long-wave spatial signal (extracted by a combination of multilooking and Gaussian filtering) and signal related to atmosphere that can be modeled (e.g. using COMET GACOS data or by correlation of phase with a DEM). We show the effect of the improved approach is significant in standard time series processing (here using LiCSBAS).