[R-sig-Geo] clustering multi band images

[Edzer Pebesma]

Laura Poggio wrote:
> However the results are quite different from what expected.
> Then I am wondering if is it that the correct way to handle clustering of
> remote sensing images.
>
> Thanks
>
> LP
>   
Laura, it's hard for us to look over your shoulder, and say something 
useful, we also don't know your expectations.

Please remember that clustering is done only on the pixel values, and 
ignores the spatial ordering of those pixels. If there's a bit of noise 
in the image, you might find quite a bit of noise in the resulting clusters.
--
Edzer