[GRASS-user] Error with sub-group while using i.gensigset

[Markus Neteler]

On Wed, Apr 14, 2010 at 7:41 PM, Luisa Peña <luisapena1979 at gmail.com> wrote:
> Greetings
> I'm freely following this tutorial
> (http://grass.osgeo.org/wiki/Image_classification) and I'm getting a odd
> error.
> I'm sending this command:
> i.gensigset trainingmap=a at landsat group=gr06 at landsat subgroup=sub
> signaturefile=sig maxsig=5

Do you really need all @landsat indications?

> and I get this:
> Finding training classes...
> 5 classes found
> Reading raster maps...
> Clustering class 1 (1753 pixels)...
> Number of subclasses is 5
> Clustering class 2 (2810 pixels)...
> Number of subclasses is 5
> Clustering class 3 (2772 pixels)...
> Number of subclasses is 5
> Clustering class 4 (2177 pixels)...
> Number of subclasses is 5
> Clustering class 5 (6060 pixels)...
> Number of subclasses is 5
> unable to create signature file sig for subgroup sub of group gr06 at landsat
> ERROR: Unable to create signature file <sig>
>
> I think it might have somethibg to do with i.group because everytime I open
> i.group the followiong is printed to the Command output:
> Invalid element 'gr03 at landsat@landsat'
> Invalid element 'gr04 at landsat@landsat'
> Invalid element 'gr05 at landsat@landsat'
> Invalid element 'gr06 at landsat@landsat'
> All these are groups that I have made
>
> If I do  group -l
> group=gr06 at landsat                                                   I get:
>
> group <gr06> references the following raster maps
> -------------
> <lsat7_2000_10 at landsat>    <lsat7_2000_20 at landsat>
> <lsat7_2000_30 at landsat>    <lsat7_2000_40 at landsat>
> <lsat7_2000_50 at landsat>
> -------------
> i.group complete.
>
> What might be wrong?

I dunno but perhaps the @landsat isn't properly stripped off in some cases.
Could you retry to make the group etc in the landsat mapset itself in
order to understand if @landsat makes the difference?

Markus