PROCESSING "LUT=.... issue

[A.Th.C. Hulst]

Frank,

On Thursday 19 July 2007 18:39:38 Frank Warmerdam wrote:
> Do you have a sense where things are going wrong?  Is it the scaling or
> the LUT application?
I think the problem lies in the LUT application. When all 3 bands end with 
255:255 then you really say that the highest value is always white. When you 
apply a max scaling then the area with the maximum value is larger and more 
apparent. 

>
> I will note that the lut management code assumes that 0 maps to 0 and
> 255 maps to 255.  So I don't think your 255:0 specifications are
> going to work.
If I don't put the 255:0 there then there's an interpolation between the last 
pair (157:0) and 255:255 which is undesirable. By putting the 255:0 there I 
prevent the code from interpolating "back up". My last pair should formally 
be 254:0 which then is appended with 255:255.

All in all I don't think that automatic catenation of pairs is a good idea 
because min and max values are fixed to black and white repectively.

Best regards,
Sander