Sentinel 1A preprocessing steps

Hi everyone,

since i am new to the processing of sar data, i got a few questions about this topic. I want to use Sentinel 1A GRD data as a byproduct for a land use/ land cover study along with the Sentinel 2A data. Until now I found out that I basically need the following steps:

1: calibration of the DNs to Sigma0, which should deliver the appropriate backscatter values (?) (Radar > Radiometric > Calibrate)

Do I need to check the “save as complex” and perform the conversion “Linear to dB”?

2: despeckle (Radar > Speckle Filtering > Single Product Speckle Filter)

3: terrain correction (Radar>Geometric > Terrain Correction > Range-Doppler Terrain

Is there any other step necessary to perform the preprocessing? Any help would be appreciated.

Thank you in advance

GRD-products are not complex (like SLC-products) so no need to save as complex. Linear-to-dB should be performed after speckle-filtering, not before it.

You might want to consider using a co-registered stack of GRD-images, in which case you can perform multi-temporal filtering for speckle reduction.

1 Like

Thank you for the advices. So I guess that my backscatter values of Sigma0 derived from the “linear to dB” conversion would correspond to the reflectance values in optical remote sensing.

I also read about the radiometric normalization that can be performed during the terrain correction. Is this a common preprocessing step?

Radiometric normalization is common and helpful in hilly areas. Radiometric terrain flattening works even better for diminishing topographic effects on SAR imagery.

Dear all,
I have some questions about SAR prepossessing steps; actually I do not know what processes should I apply on those products;

I have GRD product, I saw in below link that ‘radiometric correction’ and ‘multilooking’ have been applied before on SAR images, what is left?
In my idea;
A. Spackle filtering
B. Terrain correction (Although we do not need this step, if we are doing processing on ice)

  1. Am I right? Is there any extra steps that I have to apply and why? Like ‘apply orbit file’ and ‘resampling’?
  2. Is ‘re-sampling’ same as ‘geocoding’?
  3. I know differences between ‘beta_0’, ‘sigma_0’ and ‘gamma_0’ but I do not know when we should work with them?
  4. Why we do not use them (‘beta_0’, ‘sigma_0’ and ‘gamma_0’) in preprocessing steps?

Please guide me.

Multilooking is optional, it reduces spatial resolution while increasing radiometric resolution. Same thing with speckle filtering.
You should perform removal of thermal noise and calibration.
Since you are working on the sea surface you will need to do ellipsoid correction (instead of terrain correction)