[GRASS-user] i.group and r.patch in supervised classification

[Markus Neteler]

On Tue, May 17, 2011 at 4:01 PM, Chethan S. <chethanuniversal at gmail.com> wrote:
> Hi all!
>
> I am confused about the utility of r.patch and i.group w.r.t. image
> classification.
>
> In ERDAS for supervised classification layer stacking can be done followed
> by classification.
>
> A way to layer stack in GRASS would be r.patch (since layers are more than
> 3, I can't use r.composite). But for classifying in GRASS since i.group has
> to be used, can I use i.group to do layer stacking instead of using r.patch
> and then classify?

If you want to mosaik (compose spatially *adjacent* maps), then use r.patch.

If you want to vertically merge images in order to close holes, use r.patch.

If you want to vertically merge images in order to make a RGB composite
or HIS composite, use r.composite etc

If you want to *logically* combine a stack of maps in a group (that's only
targeting the metadata essentially), use i.group.

Most i.* commands expect a group to be defined in order to easily tell them
which images to be analysed. That's all that i.group does.

Hope this helps
Markus