[CREATE] Lens correction database

[Pablo dAngelo]

Hi all,

> From Tue, 8 Jan 2008 20:15:53 +0100
> sebastian sauer <scb at lo-res.org> wrote:
> 
> > also a side topic could be semi-automatic camera-calibration (that
> > means semi-automatic dedection of e.g. the lens parameters)
> [...]
> > ..but this stuff could be added later to lensfun, as an alternative
> > or enhancement to the lens database.
> Lens calibration is the main thing still missing in lensfun. Without it
> there's no way to extend lensfun database, except manual work with
> MathLab/Octave and alike... and it's a time-consuming thing.

Hugin is able to provide the radial distortion correction parameters,
also based with a single photo with straight lines inside. I agree that
it is probably not the easiest GUI to do just lens calibration, but I guess
its the best that is currently available...
http://wiki.panotools.org/Lens_correction_model#Determine_lens_correction

Btw. there is a tool called cam checker that uses a high quality calibration
algorithm (similar to the Matlab camera calibration toolbox).
I haven't checked their implementation in detail yet, but at least their
method of segmenting the checkerboard is not really good and probably
leads to less than perfect results, because only imprecise corner points
are found. For good results they need to be refined down to sub-pixel levels.
This might be relatively easy to add though.
http://matt.loper.org/CamChecker/CamChecker_docs/html/index.html

The advantage of that algorithm is that it also accurately determines
the focal length and not only the radial distortion parameters, if a proper
a number of images and a good distribution of checker board poses
is used.

> Some of my thoughts on this topic are on lensfun.berlios.de.

One problem with the "sheet calibration" approach is that it does
not reflect the typical shooting situation, where the objects are (relatively)
far away from the camera. The distortion is also focus dependent,
this effect is stronger if the object is very close to the camera (think A4
calibration sheet). A general rule of the thumb is to calibrate in a similar
distance as the final range.

However, it will definately better than no calibration at all. I don't believe the
vignetting can be recovered faithfully from it though.

Btw. after the most pressing issues with hugin have been fixed and a new
stable version is released, I'll look into adding an automatic
"submit calibration to lens database" feature.

ciao
  Pablo
_____________________________________________________________________
Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
http://smartsurfer.web.de/?mc=100071&distributionid=000000000066