Dear all,
I’m a little bit confused regarding the possibility to calibrate my data (S1_GRD) directly in Gamma0:
I’ve got that Gamma0 could be a very interesting output if you calibrate before Beta0 and then apply the TerrainFlattening. But I can’t read anything about the usefulness to obtain Gamma0 directly by the Calibration operation without applying TerrainFlattenig afterwards (as shown on the image above).
Thx for any help!
The Gamma0 of the calibration module is based on this formula and uses the cosine of the local incidence angle.
Source
The flattened Gamma0 of the Terrain Flattening module is the one published by David Small (@eyeinsky) and uses the integrated surface area illuminated per pixel instead of the angle as described in this study:
Small (2011): Flattening Gamma: Radiometric Terrain Correction for SAR imagery”, IEEE Transaction on Geoscience and Remote Sensing, Vol. 48, No. 8. (PDF)
The latter is more complicated to compute, but produces better results.
Dear Mr. @ABraun,
Thanks a lot for your reply.
I would have a couple of questions/considerations:
Would it be possible to say that the formula to calculate Gamma0 by the calibration module is obsolete or useless? I’ve never read on the forum about its utilization!!
If the two calculations for Gamma0 are definitely different, in order to avoid confusion, wouldn’t it be better to call the output of TerrainFlattening module with another name (e.g. GammaSmall0 or Wahtever0)?
Have a nice indoor Easter Monday
I would not say that it is obsolete, it is still widely acknowledged and applied in science and applications, but it is not very effective in areas with strong topography.
About the naming:Both are techniacally one level above Sigma0 which only uses the global incidence angle, but use different methods of estimating the impact of topography. But your point is legit - I will suggest this to the developers.
Dear @ABraun,
Thanks a lot for the clarifications.
Anyway, regarding the argument:
“Sigma0 which only uses the global incidence angle”
The formula seems to take into account the local incidence angle also for the calculation of Sigma0:
Could you please help me to understand this point?
sorry, this was confusing from my side.
Both Sigma0 and Gamma0 can be normalized against the local incidence angle.
I tried to summarize this in my work, please have a look at pages 31-33 where it is - hopefully - presented more clearly.
Dear @ABraun,
your work looks absolutely great, from the technical point of view and for the thematic applications (I’ve worked on mapping Nigerian IDP Camps)!
Thanks a lot for sharing, it will be very helpful for me to understand many points.
Best
happy to hear that!
Thanks for sharing your work @ABraun ! At last an illustrative presentation of calibration issue.
The calibration module actually uses the global (ellipsoidal) incident angle as far as I know, not the local incident angle, just as sigma0 is based on the ellipsoidal incident angle.
thank you for pointing this out. I agree, maybe my phrasing was unfortunate.