Beginner advice : DEM, choice of sentinel data, snaphu

Hello,
I am a new user of SNAP and I have difficulty understanding the big idea among the abundance of information available.
I started looking into inSAR principle, the tutorial about sentinel+SNAP, and the ones written by Luis Veci about Interferometry “TOPSAR and Stripmap”.

Now here is what I want to do : I am trying to obtain DEM of mountain (chinese ones), we were using software like Global Mapper before, but we were hoping that sentinel data could give us a better accuracy, or at least data that are RAW and not already process like ASTER GDEM (all angles and slope are smoothen on those data, so not really useful data).

So it seems to me that I just need to follow “Stripmap Interferometry with Sentinel-1 Tutorial” by Luis Veci, and select TOPO mode when exporting in Snaphu.

If it’s not at all what I should do please tell me^^. Moreover can someone help me understant the difference between those two tutorial TOPSAR and Stripmap ? And another thing, when I will obtain my coherence value, what can I consider has a good coherence for DEM ? So that I don’t start the calculation for nothing.

Thank you for your help, and for everybody writting all those tutorials.

DEM-generation with repeat-pass InSAR is very challenging due to the influence of atmosphere, ionosphere and temporal decorrelation. Is the SRTM-DEM of not high enough for you?

1 Like

Well we want to obtain a better accuracy/resolution. For example we were studying a mountain, on the DEM from ASTER or SRTM the slope seemed good, but then we had to buy DEM from Airbus (worldDEM/SPOT satellite) to obtain a better resolution, and it appeared that this nice slope is just a serie of cliff, it looks like stairs in reality. So we want to find another solution to obtain better DEM. From your point of view, it’s possible with sentinel data ? Or should I focus on other ways (I am also looking into using drones/UAV). Thank you for your help.

Could you join the URL links to the cited documents cited?

Does the standard InSAR processing chain work for you?

I’d expect that the low-frequency errors in the S1-produced DEM are much bigger than in SRTM. If you could find a way to warp the S1 DEM onto the SRTM DEM using GCPs, the result might be better locally. The TanDEM-X DEM would be many orders of magnitude better for sure.