[GRASSLIST:9965] using i.class

[maning sambale]

Dear list,

Some basic questions in using i.class module.  We have been
classifying landsat tm images from GLCF and we're using i.class to
identify training regions.
My questions are as follows:
1.  For a region of approximately 500 by 500 cells at 30 m pixel res,
how large or small should one training region be? Markus & Helen's
book said it "is important not to digitize very small areas since this
will be ignored".
2.  How many training region's should be digitized for each landcover class?
3.  Can I use the saved signature file from i.class for an i.smap
reclassification?
4.  Finally, using grass on cygwin, processes that require extensive
use of the mouse (i.rectify, digitizing, i.class) makes grass run very
slow.  Is this normal?

Cheers,

Maning

GRASS user from the Philippines, I know only one Filipino aside from
me and my students who use GRASS in the Philippines.  If there are
anybody else please let me know.

--
|---------|----------------------------------------------------------|
| __.-._  |"Ohhh. Great warrior. Wars not make one great." -Yoda     |
| '-._"7' |"Freedom is still the most radical idea of all" -N.Branden|
|  /'.-c  |Linux registered user #402901, http://counter.li.org/     |
|  |  /T  |Philippine Biodiversity Web Map                           |
| _)_/LI  |www.geocities.com/esambale/philbiodivmap/philbirds.html   |
|---------|----------------------------------------------------------|