[ptx] automatch: another attempt at semi-automatic panorama stitching

[Bruno Postle]

On Sun 04-Jan-2004 at 10:45:42 +0100, Pablo d'Angelo wrote:
> 
> > Not at all, the feature-points are mapped to a spherical surface in
> > 3d space before being analysed, but this is necessary even if the
> > eventual result is a rectilinear or cylindrical panorama.
> 
> Hmm, why don't use a spherical coordinate system? instead of 3d positions on
> a sphere. sqrt(theta^2+phi^2) would then measure the angular distance, I think.

Because I forgot all that math (and more) a *long* time ago, though
I can still remember pythagoras and basic trigonometry.

> I couldn't resist and read the sourcecode you posted. wouldn't using a
> hashmap speed up Similar_Links() considerably? probably adding a distance into
> neigbouring bins also, to simulate the margin you use with the $upper and
> $lower variables.

Yes the link matching code is extremely inefficient, it does a full
search of the entire dataset every time - A lookup table that stores
all the possible results might speed the program up 100 times.

-- 
Bruno