[GRASS-user] Supervised classification without i.class

[Daniel Victoria]

Hi Luis,

I have imported shape files in grass 6.4 using v.in.ogr. Not a big hassle.

One way to acquire training areas for the supervised classification is
to digitize them using either the tcl/tk digitizer or, what I find
easiest, the QGis digitizer. QGis is free and can create shape files.
I'm not sure (never tested it) but I believe you can even use QGis to
create Grass vector files.

Once the vectors are done, just convert them to raster, use i.gensig
to extract the signatures and then use i.maxlik to classify the image.

cheers
Daniel

On Thu, Dec 3, 2009 at 9:33 AM, Luis Lisboa <luislisboa1975 at gmail.com> wrote:
> Hello Martin thanks for your reply.
>>I am afraid not -- only non-interactive approach. Digitize areas, rasterize
>> vector areas and >use i.gensig.
> Ok. But If I try to have Linux or Mac OS machine, I could use i.class right?
>
> About your links, ehehehe unfortunely Czech is not in my language domain
> (English, Spanish, French, Catalan). But I will try to get some ideas :)
> Thanks
>
>>Import data to native format using v.out.ogr or link data using
>>v.external (some problems can occur). The OGR interface is being
>>improved in GRASS 7.
> Uhm I'm planning to use 6.4 (it has binaries). has anyone tried with success
> this operation?
> Thanks
> Best regards
> Luis Lisboa
> Universitat Pompeu Fabra, Barcelona
> _______________________________________________
> grass-user mailing list
> grass-user at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/grass-user
>
>