Relative Polarimetric Calibration - ALOS

Dear collegues,

I have checked the Relative Polarimetric Calibration steps in the given reference for ALOS and would like to confirm if all of them are already implemented at the complex radiometric calibration in SNAP:

  • correction by distortion matrix
  • compensation for Faraday Rotation
  • cross-pol symmetrisation


SNAP does not do any Faraday rotation correction and we need it for L-band.

Thanks for your answer, lveci!

I have, then, created new I and Q bands corrected by the FR angle, using the band math operator.
As example:

  • i_HH_FR = cos(FR_angle)*( cos(FR_angle)*i_HH - sin(FR_angle)*i_VH) + sin(FR_angle)*( cos(FR_angle)*i_HV - sin(FR_angle)*i_VV)

  • q_HH_FR = cos(FR_angle)*( cos(FR_angle)*q_HH - sin(FR_angle)*q_VH) + sin(FR_angle)*( cos(FR_angle)*q_HV - sin(FR_angle)*q_VV)

Although I can not assign the new bands as ‘real’ or ‘imaginary’ in order to run a decomposition or other polarimetric operator.
The SNAP does not recognize them as Scattering components.


1 Like

Did you set the ‘unit’ to real and imaginary?
You could also try not making the new band virtual and removing the original i q bands such that there is only the i q FR bands in the product.

1 Like

It has worked!!! Thanks!
Creating it NOT as “virtual band” and setting the ‘unit’ in properties.

A suggestion: You should consider enable the unit setting during the band creation.

Regards and congratulations for the efforts of the developing team

Gustavo, let us know what is needed for these corrections (reference paper or equation) and we could put into SNAP.

The equations for FR correction may be found in the section 6.2 of

OPS, the link is broken!
The report (section 6.2):
Calibration_palsar_products_v13-annotated.pdf (151.7 KB)

that is a great suggestion - thanks @gusortiz!
I always thought about the calibration of ALOS2 FullPol and how I could solve it.

Thanks also for the consideration of implementing it in SNAP, @lveci.

dear @gusortiz: How did you calculate the FR_angle variable?
I looked it up in Annex C in the PDF you mentioned but I’m not sure about the notation in SNAP.

@ABraun, there is a way to estimate the FR_Angle to each pixel as you’ve mentioned.

But, during this preliminary approach in SNAP, I used the mean FR angle for the scene provided by the ASF.

There is an important point:
when the FR angle is low (<5º) the distortion is negligible for some applications.

Please check the Freeman and Saachi (2004) article:

1 Like

thanks, I’ll have a look at it!