hugin-devs team mailing list archive
-
hugin-devs team
-
Mailing list archive
-
Message #04902
[Bug 1277770] [NEW] Outlier rejection prevents accurate optimization
Public bug reported:
I have been trying to stitch a 360 degree by ~40 degree panorama and it
always has a stitching error somewhere around the circle.
What seems to occur is it stitches everything around up until it finds
that it went too far/not far enough, and then the last set of control
point pairs then have a distance of hundreds of pixels, causing the
algorithm to ignore them entirely. Even changing the optimization
parameters to only allow changes to the misplaced photos in question
doesn't work, because the control points still count as outliers and it
does nothing.
If possible, I'd like adjustable outlier rejection so that I could force
the images into line. I ended up having to select the least-interesting
part of the 360 degree circle and split the pano there by removing the
control points.
Operating System: Linux 3.11.0-15-generic x86_64
Architecture: 64 bit
Free memory: 8018064 kiB
Hugin
Version: 2013.0.0.76c3df493921
Path to resources: /usr/share/hugin/xrc/
Path to data: /usr/share/hugin/data/
Path to user lensfun database: /home/carvac/.local/share/lensfun
Libraries
wxWidgets: 2.8.12.1
libpano13: 2.9.18
Boost: 1.53.0
Exiv2: 0.23.0
Lensfun: 0.2.7.0
** Affects: hugin
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Hugin
Developers, which is subscribed to Hugin.
https://bugs.launchpad.net/bugs/1277770
Title:
Outlier rejection prevents accurate optimization
Status in Hugin - Panorama Tools GUI:
New
Bug description:
I have been trying to stitch a 360 degree by ~40 degree panorama and
it always has a stitching error somewhere around the circle.
What seems to occur is it stitches everything around up until it finds
that it went too far/not far enough, and then the last set of control
point pairs then have a distance of hundreds of pixels, causing the
algorithm to ignore them entirely. Even changing the optimization
parameters to only allow changes to the misplaced photos in question
doesn't work, because the control points still count as outliers and
it does nothing.
If possible, I'd like adjustable outlier rejection so that I could
force the images into line. I ended up having to select the least-
interesting part of the 360 degree circle and split the pano there by
removing the control points.
Operating System: Linux 3.11.0-15-generic x86_64
Architecture: 64 bit
Free memory: 8018064 kiB
Hugin
Version: 2013.0.0.76c3df493921
Path to resources: /usr/share/hugin/xrc/
Path to data: /usr/share/hugin/data/
Path to user lensfun database: /home/carvac/.local/share/lensfun
Libraries
wxWidgets: 2.8.12.1
libpano13: 2.9.18
Boost: 1.53.0
Exiv2: 0.23.0
Lensfun: 0.2.7.0
To manage notifications about this bug go to:
https://bugs.launchpad.net/hugin/+bug/1277770/+subscriptions
Follow ups
References