360 Pano with Nikon 10.5mm fisheye shaved
RTLB
I am having a problem with the PTGUI stitching with fisheyes.
I have read that there is no definite No-Parallax-Point (NPP) for
fisheyes, so the light that comes from the peripheries would be on a
different NPP than the one coming in the middle and around. Therefore
one would always have stitching errors.
Is this correct? Because I am facing a problem where parts (usually
the higher parts) of the seams are not stitched correctly, while the
bottom and middle part is fine.
Has anyone found a way to correct this without post-processing on
photoshop?
Regards,
RT
Erik Krause
> I am having a problem with the PTGUI stitching with fisheyes.
> I have read that there is no definite No-Parallax-Point (NPP) for
> fisheyes, so the light that comes from the peripheries would be on a
> different NPP than the one coming in the middle and around.
This is true.
> Therefore one would always have stitching errors.
Not always. Only if near and far objects are in the overlap regions.
> Is this correct? Because I am facing a problem where parts (usually
> the higher parts) of the seams are not stitched correctly, while the
> bottom and middle part is fine.
If your pivot point is set inside the range the NPP moves the parallax
error should be a few pixels only. If it is more most likely something
else causes the stitching errors.
See also http://wiki.panotools.org/Parallax
> Has anyone found a way to correct this without post-processing on
> photoshop?
Optimize for lens shift, too (Optimizer tab, Horizontal shift and
Vertical shift). There is hardly a lens-camera combo where image center
and optical axis aren't offset some dozen pixels.
--
Erik Krause
http://www.erik-krause.de
PTGui Support
> Hi All,
>
> I am having a problem with the PTGUI stitching with fisheyes.
> I have read that there is no definite No-Parallax-Point (NPP) for
> fisheyes, so the light that comes from the peripheries would be on a
> different NPP than the one coming in the middle and around. Therefore
> one would always have stitching errors.
This is true but the effect is very small, in the order of a few pixels.
Usually these small stitching errors will be invisible after blending.
> Is this correct? Because I am facing a problem where parts (usually
> the higher parts) of the seams are not stitched correctly, while the
> bottom and middle part is fine.
Sounds like you have more severe parallax, you may want to check the
calibration of your panoramic head.
In general if you have stitching errors please see 4.4 in the FAQ:
http://www.ptgui.com/support.html#4_4
Joost
Hans
On Feb 26, 9:57 pm, RTLB <royt...@gmail.com> wrote:
> Hi All,
>
> I am having a problem with the PTGUI stitching with fisheyes.
> I have read that there is no definite No-Parallax-Point (NPP) for
> fisheyes, so the light that comes from the peripheries would be on a
> different NPP than the one coming in the middle and around. Therefore
> one would always have stitching errors.
>
> Is this correct? Because I am facing a problem where parts (usually
> the higher parts) of the seams are not stitched correctly, while the
> bottom and middle part is fine.
lens correction when used as shaved.
It has an extremely compressed outer area from 160 to 200 degrees.
This means that to get perfect correction for the extreme zenith you
need to place controlpoints manually
in those areas.
One tip for you. start with setting a value of -0.16 for the a
parameter in the lens settings. Also set the FOV to 190 and the
cropcircle just inside the edge.
Then optimize only for FOV +b + vertical and horisontal shift. You
have to use the advanced optimizer.
Your FOV should end up at around 186-190 degree.
If your zenith comes out perfect you can now save this as a template.
If you still have problems at zenith ad controlpoints at the zenith
and optimize with the a booked.
Hans
RTLB
and what is the best solution to fix it.
Regards.
Zeljko Soletic
Where is NPP at 4 shots, and where at 3 shots?
Erik Krause
> Where is NPP at 4 shots, and where at 3 shots?
find an animated ray tracing diagram for the Nikkor 10.5mm
Sam Rohn
- - - - - - - - - -