« Return to Thread: autopano-sift-c and "foreign" project files

Re: autopano-sift-c and "foreign" project files

by Tom Sharpless :: Rate this Message:

Thanks, Harry

The memory leak question I will definitely look into.

As for the other, my feeling is that someone else should try to make
the "bells and whistles" such as align work right. They were pretty
broken when I started doing surgery on autopano-sift, and I didn't
even try to keep them working, just to make the core CP finder
better. Some of them are poor substitutes for things the stitchers do
better, and probably should be removed, but any that still may be
useful should be fixed.

I'm not sure a segfault isn't the best way to "warn" that some
combination of options doesn't work -- better than generating bogus
results, anyhow :)

Cheers, Tom

On Jul 5, 2:00 pm, Harry van der Wolf <hvdw...@...> wrote:

> Hi Tom,
>
> When you request for sugestions for fixes, there are two:
> hugin-bugs-2795133<https://sourceforge.net/tracker/?func=detail&atid=550441&aid=2795133&...>
> autopano-sift-c has SEVERE memory leaks
> The other one is not a real bug in the tracker but one mentioned in the mail
> (I didn't check it myself yet). It's the autopano-sift-c segfault if RANSAC
> is not used and align is
> requested<http://groups.google.com/group/hugin-ptx/browse_thread/thread/57c3549...>
>
> If you find the time, inspiration and insight, it would be nice if these
> bugs were solved.
>
> Harry
>
> 2009/7/5 Tom Sharpless <TKSharpl...@...>
>
>
>
> > Hi
>
> > While looking into the feasibility of using autopano-sift-c with
> > PTAssembler, I discovered that APSC cannot load file specifications
> > from either PTAssembler or PTGui project files. This is too bad,
> > because that is the preferable way of using APSC. Not only does it
> > save retyping the file specifications, and automatically invoke the
> > stereographic mode for fisheye lenses, but also the .pto supplies
> > field of view info that APSC needs (enabling the processing of multi-
> > lens image sets, for example).
>
> > The idea was that APSC should be able to get the needed file info from
> > any PanoTools compatible script (including Hugin, PTAssembler and
> > PTGui project files). It outputs a pretty generic PT script
> > containing the control points. If it did load a script, the output
> > also contains source file specs, and is a usable stand-alone Hugin
> > project file.
>
> > The second part works for Hugin and PTAsm. Hugin 0.7 can still run
> > from an APSC output script; PTAssembler 5 can't use it as an original
> > project file, but will happily load control points from it. PTGui 8
> > won't do either.
>
> > But the first part is not working because both PTGui and PTAsm project
> > files now store the file names in "comment lines" of the form #-
> > imgfile width height "pathname" (hugin.pto's still put those data on
> > the i or o lines, as in PanoTools). So APSC can't see any files in
> > those scripts.
>
> > If nobody has an objection, I intend to patch APSC so that it can load
> > PTAssembler and PTGui scripts too. Please post any other suggestions
> > for improvements or fixes; I might do those at the same time.
>
> > Regards, Tom

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@...
To unsubscribe from this group, send email to hugin-ptx-unsubscribe@...
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~----------~----~----~----~------~----~------~--~---

« Return to Thread: autopano-sift-c and "foreign" project files