[GRASS-user] i.maxlik yellow map

[Moritz Lennert]

On 07/06/18 11:00, Paul Shapley wrote:
> Hi Users,
> 
> I am re-posting this to see if anyone else may have some suggestions. My 
> thanks to Veronica who suggested checking 'g.region' but this appears to 
> be fine. I've also tried the same on Grass 7.2.2/7.4 and nightly but all 
> produce the same result.
> 
> signature file content:-
> 
> #
> #Bare_peat
> 2667621
> 60.4424 75.4845 87.2587
> 716.802
> 831.829 1086.7
> 869.829 1105.97 1179.92
> 0 0 0
> 
> ------------------------------------------------------------------------------------------------------
> 
> Original message:-
> 
> I have created a signature file in the g.gui.iclass (supervised 
> classification module) and it looks correct in 'preview' and when i 
> inspected the 'sig' file. When the signature is applied using 'i.maxlik' 
> the result is a solid yellow colour as a layer. What could have gone wrong?

Do you only have one class ? Then it is normal that all pixels will be 
in the same class. You could ask i.maxlik to output the reject map and 
then define a threshold in that map as a mask to show those areas whose 
probability of being in the class is sufficiently high.

Moritz