[GRASS-dev] Landsat 5TM pre-processing - histogram matching - problem

[Nikos Alexandris]

* joanna mardas <joanna.mardas at wp.pl> [2015-05-05 00:19:02 +0200]:

>    Hello!
> 
>    You both are my heros! Thank you Moritz and Nikos!
> 
>    r.recode worked :D I did i.histo.match on two test images and it looks
>    fine.

If you go all the way from DN to Top-of-Atmosphere reflectances (via
i.landsat.toar), then to Top-of-Canopy (via i.atcorr), you'll have
floating point values ranging in [0, 1.0].  If you recode this back to 8-bit,
you should consider whether there is an important loss of the radiometric resolution.

So, recoding to another range is different than converting to integer
numbers and trying to preserve the range. 


>    So now the rest of bands, and then band composite and supervised
>    classification (I've found nice tutorial on youtube, so I'm gonna follow
>    it, or I will try semi-automatic classification plugin for qgis).
> 
>    Nikos, you've wrote "Each band is one single image. No matter its bitness,
>    it is "normal" to appear "mono-chromatic" -- its "normal" to apply a grey
>    scale." I know that grey scale is normal, but I didn't have, let's say "50
>    shades of grey" ;) but only one shade, one color-no scale of colors.

Thanks for the details Joanna. I now understand better. After each
processing you can check the range of your image(s) via `r.info -r` and
or r.univar for basic descriptive stats.  If all looks fine (min value,
max value, mean and standard deviation show clearly there is useful
information in the queried image(s)), then it's only a color-table
issue.  You should only need to apply the correct color table via
r.colors.

>    Forunately, Moritz was right, thanks a lot! Now it looks normal.

Sure. But consider what is your aim and if you are ok giving away
some portion of the radiometric details.

>    And you wrote also that "the correct way is to not use the Digital Numbers
>    directly. Rather, convert to reflectance, then do whatever you have to."
>    So application of i.landsat.toar is correct, right?

Yes.
> 
>    Sorry for all stupid/basic questions but this is my first serious time
>    with grass and landsat image processing.

If you do request (from me, at least) to accept apologies, then accept
my apologies for my infinite silly questions I have asked in this list
:-)

Nikos


ps-  Please try to keep some flow-of-text from top-to-bottom when
replying in the list.