[GRASSLIST:149] problem in supervised classification

[Uttam Kumar]

Hi GRASS USERS,
               I wanted to do supervised classification.For that I had 
several training data sets-namely builtup, agriculture,waste,water etc in 
vector form.I digitized those training data through v.digit and labelled 
them as lines and gave them category values.Then I converted them to raster 
form through v.to.rast command.Finally I executed the querry after running 
the command r.mapcalc as
<output_map>=if((builtup_map==1) || (agriculture_map==2).....).The combined 
raster map was obtained as <output_map> having all the categories I named 
during digitizing (category 1,category 2,category 3..). The problem is that 
the pixels in the raster map consists of only one color representing all the 
categories instead of different colors for different categories which can be 
seen when one displays the raster images separatly for each category in a 
separate window.(eg. blue for water, yellow for agriculture, red for 
builtup...).The resultant file after combining all the training data sets 
was put as a signature file for the command i.gensig.The resulting output 
was used as a input for i.maxlik.What I get in last is a raster image of 
only one colour and other categories are not seen.Yet they can be seen while 
running d.legend. My problem is how to use available training data sets in 
GRASS through i.gensig or i.class for using in i.maxlik for performing 
supervised classification.
Please help

regards,

uttam.

_________________________________________________________________
Beat the heat. Chill out in Himalayas. 
http://server1.msn.co.in/sp03/summerfun/index.asp Mercury Rising contest.