Hi Bruno, thank you very much for looking at my data.

 

I would try deleting control points that are not near the likely seam  
position. This seems counter intuitive, but these pixels will not end up in 
your final image.

 

That sounds logical to me, even if I don't understand your reasoning. In my 
words, I would have said that it's because the fish-eye distortion is extreme 
at the edges, making reprojection very unreliable. Or am I missing something?

 

... or create lots of vertical control points to properly calibrate the lens 
distortion (I would create the vertical points).

 

Should the vertical control points only be located in the overlap area or also 
outside it?

 

I would also delete any close control points, ie. The floor might look  like a 
great calibration source, but using it will amplify any parallax  problems 
caused by the slight offset between lenses.

 

Good point, thank you.

 

Also, these circular fisheye images are very unlikely to be central in  the 
frame, so you need to optimise d & e parameters separately for  each side.

 

What do you mean by "separately for each side"?

 

Thank you very much for your valuable advice.

 

Robert

 

 

 

 

Von: [email protected] [mailto:[email protected]] Im Auftrag 
von Bruno Postle
Gesendet: Sonntag, 23. November 2025 09:00
An: hugin and other free panoramic software <[email protected]>
Betreff: Re: [hugin-ptx] Sharp discontinuities in the 360° spherical panorama

 

Hi Robert, I would try deleting control points that are not near the likely 
seam position. This seems counter intuitive, but these pixels will not end up 
in your final image.

 

Important, if you do this the optimiser can find all sorts of crazy solutions 
for the a, b & c lens distortion parameters. To prevent this you need to 
either: only optimise angle of view, or create lots of vertical control points 
to properly calibrate the lens distortion (I would create the vertical points).

 

I would also delete any close control points, ie. The floor might look like a 
great calibration source, but using it will amplify any parallax problems 
caused by the slight offset between lenses.

 

Also, these circular fisheye images are very unlikely to be central in the 
frame, so you need to optimise d & e parameters separately for each side.

 

-- 

Bruno

 

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/hugin-ptx/002e01dc5cb6%242e11e440%248a35acc0%24%40gmail.com.

Reply via email to