Sentinel-1 Processing Steps?


can anyone tell me which pre-processing steps are applied on the Sentinel-1 products?

As they are provided in gamma nought, at least calibration must be involved. In the documentation I saw that thermal noise is removed as well ( What else? In which order? With which parameters?

I’m trying to figure out if the Sentinel Hub can save me my processing routine or at least parts of it…


1 Like

Correct, calibration and thermal noise reduction are applied. This is done first and at the same time. After that the geolocation processing follows, either orthorectified or non-orthorectified, depending on you settings.
Note that as of this writing speckle filtering and terrain flattening are not supported.
I hope this helps

Thank you for the response!
I have some follow up questions:

  • Is some kind of resempling or coregistration involved?
  • Will terrain flattening be supported in the future or why wasn’t it included yet?
  • During orthorectification - what kind of DEM is used?
  • Are the original orbit state vectores used or is the updated orbit file applied? Why so? Is this relevant for this processing chain anyway?


I can respond to some of these. @marko.repse can probably answer to the rest.
-there is no coregistration involved
-there is averaging involved on lower scales. If you use it on natural scale, there is no resampling.
-there is no terrain flattening and there are no mid-term plans to do it
-we use MapZen’s DEM (

We use the original orbit state vectors as they are sufficient for GRD use. Of course this only applies to orthorectified imagery; for non-orthorectified they are not needed.

Is the sigma0 calibrated Sentinel-1 data available from sentinelhub wms service?

We currently support only gamma0.

Actually you can get sigma0, you can set this in the layer settings in your dashboard (

It is great! And thanks for the update! I actually saw it yesterday and wondered whether it’s implemented.