I am working on landslide susceptibility assessment for Uzbekistan and need better DEM for analysis. If someone could give any idea about it would be very useful.
If you are working in arid regions with limited vegetation, generating a DEM from Sentinel-1 images could work quite well. Resolutions of around 12 meters are feasible then.
But if you are working on whole Uzbekistan this is not an ideal solution because of the limited coverage of single S1 scenes. Or do you only work in a smaller part or catchment?
Generation of DEMs from SAR data requires quite a lot of computing capacity. Besides that, it is not an easy task if you are unexperienced with SAR data because it involves a lot of parameters and decisions which have to be considered well-grounded.
Breaking it down into managable pieces, DEMs can be created by:
Downloading an image pair from the same track (from the Copernicus Science Hub or ASF Vertex, for example),
Co-registering both scenes (overlays the two images at the sub-pixel level),
Calculation of an inferferogram (stores phase information) and coherence layer (as an indicator of the quality which can be achieved),
Optionally: Filtering or multi-looking the interferogram,
Unwrapping the interferogram (converts phase cycles to continuous information),
Convert unwrapped phase to elevation (converts phasse to absolute values, meters in this case).
ESA has published nice documents on InSAR, especially for beginners: