> Apologies if this is a duplicate. I posted 24 hours ago (via Google)
> and it hasn't appeared yet. Second time lucky?
> I have to find the local maxima in an image surface. Is there a more
> elegant/efficient method than simply comparing each pixel to its eight
Use a general-purpose optimiser (simplex, downhill search, Powell, etc.).
Assuming your data has some patterns (rather than just random numbers),
you'll find it rather quickly.
Global maximum/ma is a different issue because you are then not guaranteed
to find the solution, unless you scan the entire image surface.