PolSARpro ALOS2 FDB analysis

Hello, I am working with ALOS2 FBD 1.1 scenes HH HV, and I would like to understand

  1. what kind of analysis can I do with PolSARpro having only 2 polarizations: I came across a Cloud paper ‘THE DUAL POLARISATION ENTROPY/ALPHA DECOMPOSITION:
    A PALSAR CASE STUDY’ that employes similar data, but I was not able to find this procedure implemented in Polsarpro Biomass or understand how to implement it
  2. is there any tutorial/help to understand the steps/analysis that can be done once the data is imported and extracted as Sxx/Sxy and the C2 matrix created? What can be done next? As most of the processing opportunities are not working, because is not fullpol (no S2)…

we discussed in here to what extent dual pol entropy/alpha decompositions make sense: H-Alpha Plane Problems

To my understanding, it is a technically possible, but kind of counterproductive because it disregards the actual idea of quad-pol decompositions.

You can follow along steps 1-3 to prepare the data for the processing: Calibration in PolSARpro problem
Or in short form: PolSARpro_v5.0-problem with main input directory
and here: Help installing PolSARpro software step by step on detailed on the windows 7 system
Also, the steps are discussed here: PolSARpro v.5.1 unzip problem

The DualPol Alpha/Entropy classification is also featured in SNAP, by the way.

PolSARpro offers an immense multitude of polarimetric functions and modules, far more than SNAP, but you have to know at least what you want. Polarimetry has a strong mathematical and theoretical component, it is far more than executing modules and getting results. This makes it kind of hard to access.

You can start with the PolSARap Showcases (under Education)

Thanks a lot for your prompt reply and explanation! According to the mentioned paper, that gives a different result with respect to the H-Alpha Plane Problems link you indicated, the dual pol decomposition is useful in classification problems.So, it is still unclear as these seems two different opinion from experts. I will check however in SNAP the implementation, thanks.

With respect to calibration, being dual pol this does not work (no calibration allowed as no S2).

With respect to objectives, I am sorry I did not explain myself properly. I know polarimetric possibility and theory, even if I am not an expert. I am not executing modules, instead I am performing an empirical investigation and try to understand is there is a way to isolate the volumetric information content, for biomass estimation and forest damage classification.
The dualpol data limit the polarimetric analysis and I would like to be sure there is nothing I missed.

Just to ask your support again: with dualpol data, are you suggesting to move to SNAP?

Or are you saying I made a mistake in the procedure in polsarpro?
When I follow your indications I get stucked as with dual pol I can only have Sxx and Sxy and thus the steps you described in the links kindly provided are not applicable

Thanks for additional suggestions!

ps: the PolSARap showcase is again for quadpol data… my questions are referred to dualpol

I checked it and saw that too, sorry. But you can run the polarimetric dual decomposition in PolSAR pro on the imported data

It is probably easier to do it in SNAP (you can select complex output in the radiometric calibration)

I thought there would be one which could be executed on dual-pol data as well, but now I’m no longer sure if this is true.
About the applicability of polarimetric methods for dua-pol data: I’m not an expert either, sorry. You can try to derive several parameters, such as Entropy, Anisotropy (also some in the Polarimetric Parameters operator in SNAP) and see if they highlight what you want to extract from the images, maybe you have measurements and can correlate the different parameters with empirical data to build a regression or something.
But from a methodological perspective, the concepts of polarimetry are nearly not applicable to dual-pol data (especially cross-pol), I’m afraid. I visited the ESA PolSAR course in 2019 and this question was asked quite frequently, because Sentinel-1 is free and also only offers two polarizations. But most of the lecturers responded that dual-pol data is mostly not applicable here.

Thank you a lot, this is helpful. In fact I can relate the polarimetric features with empirical data and build a regression, to see if I can improve the discrimination capability in selected conditions. I also did the Polarimetry course, but years ago. The paper I mentioned from Cloud is interesting and open a possibility of HH HV polarimetric use…and he is definitively an expert! But is true that others state that it does not help. Onestly I am not an expert enough to understand if the methodology behind the paper is ok and I do not understand also the two different opinion. I this doubt I guess that it worth a trial!
Thanks again for your support

1 Like

PS: I found an additional paper that focuses on forests…
Cloude, S. R. (2009, January). Dual-versus quad-pol: a new test statistic for radar polarimetry. In Proceedings of the PolInSAR Conference—ESAESRIN, Frascati, Italy (pp. 26-30).