[GRASS-user] Using i.atcorr on SPOT4 and SPOT5 imagery

[Daniel Victoria]

Hi Markus,

Yes, 6S was writen by Vermote/Nasa but what I meant is that the
version Christo converted to C was not the one by Vermote. I believe
the NASA version works for point locations and a lookup table was
needed, so Mauro Homem Antunes, from INPE, wrote another version that
accepted images as inputs. That's the code that Christo converted.

This is explained in the main.cpp file:
---------------------------------------------------------
This program has been rewriten in C/C++ from the fortran source found
on http://www.ltid.inpe.br/dsr/mauro/6s/index.html. This code is
provided as is, with no implied warranty and under the conditions
and restraint as placed on it by previous authors.
---------------------------------------------------------

Also, from the main.cpp
-------------------------------------------------------------
/*
438	       " Incorporated into Grass by Christo A. Zietsman, January 2003.\n"
439	       " Converted from Fortran to C by Christo A. Zietsman,
November 2002.\n\n"
440	       " Adapted by Mauro A. Homem Antunes for atmopheric corrections of\n"
441	       " remotely sensed images in raw format (.RAW) of 8 bits.\n"
442	       " April 4, 2001.\n\n"
443	       " Please refer to the following paper and acknowledge the
authors of\n"
444	       " the model:\n"
445	       " Vermote, E.F., Tanre, D., Deuze, J.L., Herman, M., and
Morcrette,\n"
446	       "    J.J., (1997), Second simulation of the satellite signal in\n"
447	       "    the solar spectrum, 6S: An overview., IEEE Trans. Geosc.\n"
448	       "    and Remote Sens. 35(3):675-686.\n"
449	       " The code is provided as is and is not to be sold. See notes on\n"
450	       " http://loasys.univ-lille1.fr/informatique/sixs_gb.html\n"
451	       " http://www.ltid.inpe.br/dsr/mauro/6s/index.html\n"
452	       " and on http://www.cs.sun.ac.za/~caz/index.html\n"; */
-------------------------------------------------------------------------------------------------

Another thing is that the version in i.atcorr is based of  "6s version
4.2b " (explained in 6s.cpp), older than the current 6S used at NASA
and MODIS processing. But I'm not sure what are the differences.

Cheers
Daniel


On Wed, Jul 18, 2012 at 8:37 AM, Markus Neteler <neteler at osgeo.org> wrote:
> On Wed, Jul 18, 2012 at 1:02 AM, Daniel Victoria
> <daniel.victoria at gmail.com> wrote:
>> From what I recall, i.atcorr was based off a fortran code writen by
> ...
>
> Fortran 6S was written by Vermote/NASA (http://6s.ltdri.org).
> This code was then rewritten by Christo Zietsman in South Africa to C
> and updated by various developers.
>
>> That original Fortran program expects the input
>> images to be 8 bit and produces another 8bit image as a result. I
>> believe that's why the grass i.atcorr has both the iscl and oscl
>> options.
>>
>> In my opinion, these input and output scaling parameters confuses
>> things a little. I'm not sure how the command works internally and if
>> these parameters are needed but, I have a feeling they are necessary
>> if we work with images with more than 8 bits.
>
> A user would likely prefer that the scaling be done automatically...
> Seems we have to study the source code...
>
> Markus