auto-finetune

[Mike Runge]

Hi,

I did a lot of work last weekend using hugin-20031001.tar.gz. I stitched a
multirow pano with 2 rows (22 photos, 11 photos each) - so it was mainly
picking points (~ 4-5 points per pair, overall about ~ 140 points).

hugin was stable and auto-finetune works well. I used the defaults for region
size. I would say auto-finetune(aft) had a quote in finding the correct
matching point about 70%. In some cases aft used permanently a wrong point -
placing the cursor different didn't helped in that cases. In 3 or 4 cases aft
was not able  to find a 'matching' point at all.

I played a lot with picking the second point in a way, that it should be
easier for recognition (nothing else similar than THE point in region,
picking exactly the middle, etc.)

In 60% of the cases, where aft fails in first instance, picking sligtly
different in a second try works. In the other 40% I had to give up matching
theses points in 'fit-to-screen-view'. Picking another point or working in
100% view is the woraround. Nearly all the times aft works fine in 100%
viewsize - but this mode is uncomfortable to work with (scrolling, overview
of picked point', etc.)
My pictures have a size of 1536x2048 (portrait).   

My idea (beneath make the algorithm better ;-) ) how handle that from a user
standpoint (weather it's feasable or not) would be the following:

#idea1
- picking a pair like today 
- if aft visible fails, just retry second point setting applying aft again
but just picking the second point again (maybe sligtly different position)
- apply 'ADD' if you think the point matches
That's similar to PTOpenGUI without having the auto add point toggle ON, but
using the nice aft feature

#idea2
- picking a pair like today
- having 2 smaller windows with different zoom factor (like PTOpenGui) to
control matching of points.
- points are auto added like today
- if aft visible fails, just pick again the second point in one of the small
windows.
- picking points in the smaller windows does an update instead of addition.

Ok, my ideas are driven from my work with PTOpenGui ;-) Starting with
automatically created points would be much smarter (and maybe much more
interestiong to solve for programmers ;-) ).

Optimizer fails completely in my example, as well as the preview fails :-( I
got no values neithes a message from optimizer. It runs a few seconds and
nothing I got back?!

BUT:
I was able to load the generated pto-file into PTOpenGui and it performed
well with that very exact points. So hugin did a good job helping me to pick
all the points :-)

Just to give you an idea, how the pictures i worked with look like:
http://www.trozzreaxxion.net/ut/gallery/multirowpano/abd
It's not ready, I need to solve brigthness problems and water drops. Just a
good test for aft.

best, mike 
http://www.trozzreaxxion.net