Hallöchen!
[I corrected the language in my last sentence in the quotation.]
Fabien Castan writes:
>>> If we use a checkerboard for calibration, we compute both the
>>> focal length and the distortion, so the result will be correct
>>> because this relationship is independent from the focus.
>>
>> If I understand you correctly, you assume that a lens at 200mm
>> nominal and close focus, which actually has only 135mm, exhibits
>> the same distortion as the same lens at 135mm nominal and focus
>> at infinity?
>
> Yes
1. I don't think so. The lens arrangement within the lens is very
different in both setups. It would be a big coincidence if
distortion was the same.
2. You cannot achieve a measurement for focus at infinity at
maximal focal length with checkerboards then.
3. The current method is easy enough: Take pictures of buildings,
upload them.
>> Are you interested if we propose a pull request with a new
>> calibration binary (with OpenCV optional at compile time)?
>>
>> I personally no. I do not believe in checkerboard calibrations
>> for reasons I have stated earlier.
>
> It could be nice to validate with tests instead of just
> "believing", isn't it?
You cannot validate because this would mean to test all lenses. You
can only invalidate by showing that it doesn't work for one lens. I
want a reliable method, and this means to test the lens with the
settings with which it is later used.
Tschö,
Torsten.
--
Torsten Bronger Jabber ID: [email protected]
------------------------------------------------------------------------------
_______________________________________________
Lensfun-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/lensfun-users
