optimizing

Pablo d'Angelo pablo at mathematik.uni-ulm.de
Thu Aug 21 14:07:29 BST 2003


Hi,

Kai-Uwe Behrmann schrieb am Donnerstag, den 21. August 2003:

> Hi,
> what would You think about the right place for an optimize class. It shall
> simply contain the different kind of setting strategies wich are
> selectable from PanoPanel (4th tab). Later it can be extended by some
> analys routines, checking v and d,e and the cp distances.

Hmm, why don't we integrate that into the OptimizerFrame I added
recently. It should probably get a toolbar button instead of hiding it
only in the view menu.

Integrating checks etc. there should be quite simple, because the
OptimizerFrame gets the optimizer output before it is applied.

But which variables to optimize also depends if one uses
horziontal/vertical control points, or fixes a reference image.
A combo box with some prefined settings might be useful, and the first
step.

I haven't played enough with the optimizer to know excatly what should
be added there. I usually use horizontal/vertical control points, if
possible, set a couple more control points than variables to optimize
and then optimize all position variables, except for a single yaw value,
in one step.

I've never played with the distortion parameters extensively, so I'm
not sure which cominations make sense, probably a mode for optimizing
d and e is also helpful for scans. (I have to add f and g sometime as well ;).

I think the PanoPanel has only an indirect connection to the
optimizer (the optimizer is influenced by the final projection),
and should not contain other optimize functionality, that is complicated
enough to deserve a own window or tab.

ciao
  Pablo
--
http://wurm.wohnheim.uni-ulm.de/~redman/
Please use PGP


More information about the ptX mailing list