[ptx] stitching multirows using hugin and autopano

[Mike Runge]

Hi Ian,

I think we are talking about the same mysterium.
I just switched over from hugin_2004_03_25-08_41_win32.zip to
hugin_2004_07_16-08_33_win32.zip. I also observed, that optimising this 7x3
multirow didn't worked with the old version, but worked nice with the new
one. I was a bit confused from my observations because sometimes I had good
optimisation results without doing anything, sometimes I had to add much help
in form of predefining some values. Yesterday I ended with the conclusion,
that the latest hugin was able to optimize my 7x3 from scratch without any
help. Today I tried to verify that, but now it was needed to specify values
for pitch at least.

I cannot tell why this is going on. But if you have similar problems to drive
your optimisation into the right direction, predefining some values helps
:-)

Thanks for your reply. I was half the way down seeing violet mices ;-) Good
to know, that somebody had observed the same strange behavior. I could not
belive that there is a difference between the hugin versions because both use
PTOptimiser and the pano12.dll. The missing t-value is a good explanation -
thanks.

best, mike



 
On 7/21/2004, "Ian Sydenham" <ian at sydenham.name> wrote:

>Mike Runge wrote:
>> Just running pairwise optim. results in a very strange result. There was
>> nothing about using controlpoints distances to fiddle out some misplaced
>> points. 80% of the controlpoints had delta values higher than 100, some
of
>> them reported a value of more than 1000! Average after optimisation was
about
>> 70.
>I am not sure if this is similar to your problem, but I did find that
>using an earlier hugin snapshot with the .pto file generated from
>autopano I got very similar crazy optimisations. The problem was that
>hugin was not recognising the control poit pairs as "normal" and I had
>to either manually add t0 to the end of all the control point pair
>definitions in the .pto file, or manually change the control point
>definition to "normal" inside hugin (Alignment cloumn on control
>points tab). Once I did this the optimisations worked well.
>
>regards,
>
>Ian Sydenham
>
>
>