i tried to generate a DEM over an area following the tutorial inside this platform (and also other platforms); steps are described below.
Imagine: Sentinel 1 SB
- coregistration DEM SRTM 1 sec
- interferogram formation
- sentinel debrust
- goldstein phase filtering
- multi-looking (n. of looks: 6, GR square pixel )
- export and import in SNAPHU ( 16 tiles 8 core to elaborate, DEFO mode)
- Phase to elevation
- Range doppler Terrain correction DEM SRTM 1 sec
- export to Tiff
But when i look the result in Qgis i have values from -700 to +2000; maximum value of +2000 could be correct for the mountain area but the -700 is absolutly wrong because the minimum land elevation is about -10 m; this is due to the reference system for elevation (geoidic or mean sea level or distance from satellite) or i miss something?
Thanks a lot
For DEM-generation the shorter the temporal baseline the better. So 6/12 days in the case of S1. Be aware that the errors caused by atmosphere will seriously degrade the quality of your DEM with respect to SRTM and other official DEM products.
In addition, please consider to change the SNAPHU configuration from DEFO mode to TOPO mode, as you want to get a DEM out of your processing.
The DEFO mode is used when analysing the deformation due to Earthquakes or Volcanoes eruptions.
I hope this helps.
Thanks, i manage to produce a more accurate DEM, i remove the flat earth correction ( downgrade the polynomial to 1) and apply the TOPO mode, add the thermal correction of S1; this is the result but the problem of elevation remain.
I also notice that DEM starts with a quota of +20 + 30 m, my AOI has an elevation of -10 +10 m (plain), maybe this could be also a problem?
SRTM 1sec same area
result from SNAP
I think the polynomial of 1 for the orbitf file causes a ramp in your image from the edges towards the middle. Any reason why you reduced it to 1?
sorry, i’ll explain better, the polynomial i reduced is not the orbit one but is the flat earth polynomial degree inside the interferogram formation (i read inside the help that for plains it suits well)
the orbit polynomial degree was set to 3