[daip] curious OGEOM problem

[Eric Greisen]

Lawrence Rudnick wrote:
> I'm having a curious OGEOM problem, currently working in 31DEC08, 
> although I suspect this may be long-standing.
> 
> I can send you the images, detailed inputs, etc, if that would be useful.
> 
> I have a 512x512 image, 15" cell size.  imh on original shows ref
> pixels 256.5, 256.5
> 
> I run OGEOM with everything set to default and aparm(4)=0.25 .
> The coordinate values at ref. pixel stay the same (as they should),
> but the ref. pixels have now changed to 258, 255 .  The shift in
> reference pixels depends on the size of aparm(4) .
> 
> I need the output maps to have the same centers as the original
> maps (in pixels), so this is a problem.  Any advice welcome, and
> let me know if you need to actually see maps, etc.  I've only tried
> it on maps with TAN projection, not sure this is relevant.

The idea is that the expansion occurs about NX/2, NY/2 + 1 and so there 
should be some change of reference pixel.  But the code did it wrong, 
dividing by the scaling instead of multiplying.  I have fixed that but 
also added code so that the ref pixel does not change if the input and 
output images are the same size other than by the shifts (APARM of 1 and 2).

Tomorrow's MNJ will update you.

Eric Greisen