Synthetic aperture radar for DEM generation in snow-covered mountain terrain

Abstract

Digital elevation model (DEM) generation has mainly been based on optical imagery and photogrammetric techniques. However, in recent years there has been a growing interest in the use of synthetic aperture radar (SAR) for this purpose. It is mainly two techniques that are used, SAR interferometry and stereoSAR. We have studied the influence of the snow cover on the accuracy of SAR derived DEMs. Three different Interferometric ERS1/2 tandem acquisitions are investigated, a dry-snow case, a wet-snow case and a no-snow case. It is found that with wet-snow the InSAR DEM has lowest accuracy. The dry-snow DEM is less accurate than the no-snow DEM, even though the coherence Is higher. This is explained by a redistribution of dry snow that can degrade the accuracy of interferometric DEM but still maintain high coherence. Six different Radamat s2-s7 stereo pair combinations covering dry-snow/dry-snow, wet-snowtwet-snow, partly-snow/partly-snow, no-snow/no-snow, dry-snow/no-snow and wet-snow/no-snow situations are considered. With stereoSAR It is observed that the type of snow cover has no significant Influence on the DEM accuracy. This is even the case for dry-snow/no-snow and wet-snow/no-snow stereopairs, even though these DEMs are less accurate than those originating from data with the same snow cover.