[GRASS-user] i.landsat.rgb

[Markus Neteler]

On Fri, Aug 11, 2006 at 12:44:54PM -0500, Zenon Panoussis wrote:
> 
> I am confused. http://mpa.itc.it/markus/grass61/demos/i.landsat.rgb
> doesn't work in grass60, it calls r.univar.sh with -par, which
> r.univar.sh doesn't understand. Nothing strange so far,
> http://mpa.itc.it/markus/grass61/demos/ says clearly "grass 6.1".

this is right.

> So I grab grass61 and its r.univar.sh does understand -par, but
> i.landsat.rgb (the one from the web) breaks anyway because of

Note that i.landsat.rgb *in* grass61 is compliant with grass61.
The web version was a preliminary version which I have removed
now to avoid confusion (I didn't remember that it was there).

> EXT=nice. i.landsat.rgb is also in the 6.1 source and it doesn't
> use EXT=nice, but it's not at all the same as the one on the web
> and it doesn't use r.composite and the documentation *cough, cough*
> and when I try it it does nothing except break with "line 94, 100,
> 110: [: 1: unary operator expected".

Not sure what you mean:
You say that the current i.landsat.rgb *in* grass61 is broken?
I used it recently without problems.

The EXT parameter does not exist in the current version.

> By now I have no clue what I should feed to which version of
> i.landsat.rgb (the .green, .red, .blue tiles created by
> i.fusion.brovey or the raw _10, _20, _30 tiles imported with
> r.in.gdal?) and what is supposed to come out of each version.
> 
> Could someone please explain how to make an enhanced true colour
> tiff from a set of raw landsat files?

Forget about the web downloaded script.
Use the one which comes with grass61.

Then simply:
i.landsat.rgb red=tm.3 green=tm.2 blue=tm.1

hope this helps,

 Markus