i.maxlik output

[Kelly Elder]

> Lesley-Ann Dupigny says :

> ""I am in the process of doing supervised classification on an AVHRR image
> using i.class and i.maxlik. After a few attempts, my classified image now
> resembles the original visible bands...""
> 
>> I might be wrong, but, if your classified image looks like the original,
>> isn't it totally useless?

>> Agus

you are wrong, there are hundreds (thousands?) of cases where the original
images looks very much like the classified image.  the difference is that
classified image is now quantitatively explicit in your perceived
relationship, rather than just qualitatively similar.

kelly