Hi,
 
I am trying to create a more robust calibration for my AR system. I have downloaded and used the Matlab Calibration Toolbox ( http://www.vision.caltech.edu/bouguetj/calib_doc/). I have calibrated my bullet camera using 6 screenshot images and I have calculated distortion. The Matlab Toolbox generated calibration files using either: 
1) Willson-Heikkil, or 
2) Zhang methods. 
 
Of course these formats are different from ARToolKit calibration file format. I was wondering if anybody knows how I could use the Willson-Heikkil or Zhang formats instead of the ARToolKit calibration software. Has anybody managed to create a robust calibration by intergrating the above? If not is there another way that I can easily do a robust calibration?

Thanks.
 
Fotis Liarokapis

Hi,
 
I am trying to create a more robust calibration for my AR system. I have downloaded and used the Matlab Calibration Toolbox (http://www.vision.caltech.edu/bouguetj/calib_doc/). I have calibrated my bullet camera using 6 screenshot images and I have calculated distortion. The Matlab Toolbox generated calibration files using either: 
1) Willson-Heikkil, or 
2) Zhang methods. 
 
Of course these formats are different from ARToolKit calibration file format. I was wondering if anybody knows how I could use the Willson-Heikkil or Zhang formats instead of the ARToolKit calibration software. Has anybody managed to create a robust calibration by intergrating the above? If not is there another way that I can easily do a robust calibration?

Thanks.
 
Fotis Liarokapis

Hi Fotis,

I have used a similar approach to get a better (easier to use) calibration
for the ARToolKit. You could save the computed parameters as an ARParam data
file or hard code them inside your program. This works fine for intrinsic
camera parameters. However, I have not found a direct conversion for the
lens distortion parameters. Have a look at Brendan Hannigan's answer to this
question (3rd of march 2003) where he suggests to comment the body of the
'ideal2observed' and 'observed2ideal' functions, so that lens distortion is
not used. Afterwards, you can just compensate lens distortion using your own
method.

Cheers,

Fivos

-----Message d'origine-----
De : owner-artoolkit@h ..................
[mailto:owner-artoolkit@h ..................]De la part de Fotis Liarokapis
Envoyé : mardi 15 avril 2003 11:52
À : artoolkit@h ..................
Objet : alternative camera calibration methods for ARToolKit?


Hi,

I am trying to create a more robust calibration for my AR system. I have
downloaded and used the Matlab Calibration Toolbox
(http://www.vision.caltech.edu/bouguetj/calib_doc/). I have calibrated my
bullet camera using 6 screenshot images and I have calculated distortion.
The Matlab Toolbox generated calibration files using either:
1) Willson-Heikkil, or
2) Zhang methods.

Of course these formats are different from ARToolKit calibration file
format. I was wondering if anybody knows how I could use the Willson-Heikkil
or Zhang formats instead of the ARToolKit calibration software. Has anybody
managed to create a robust calibration by intergrating the above? If not is
there another way that I can easily do a robust calibration?

Thanks.

Fotis Liarokapis

Hi Fotis,

I have used a similar approach to get a better (easier to use) calibration
for the ARToolKit. You could save the computed parameters as an ARParam data
file or hard code them inside your program. This works fine for intrinsic
camera parameters. However, I have not found a direct conversion for the
lens distortion parameters. Have a look at Brendan Hannigan's answer to this
question (3rd of march 2003) where he suggests to comment the body of the
'ideal2observed' and 'observed2ideal' functions, so that lens distortion is
not used. Afterwards, you can just compensate lens distortion using your own
method.

Cheers,

Fivos

-----Message d'origine-----
De : owner-artoolkit@h ..................
[mailto:owner-artoolkit@h ..................]De la part de Fotis Liarokapis
Envoyé : mardi 15 avril 2003 11:52
À : artoolkit@h ..................
Objet : alternative camera calibration methods for ARToolKit?


Hi,

I am trying to create a more robust calibration for my AR system. I have
downloaded and used the Matlab Calibration Toolbox
(http://www.vision.caltech.edu/bouguetj/calib_doc/). I have calibrated my
bullet camera using 6 screenshot images and I have calculated distortion.
The Matlab Toolbox generated calibration files using either:
1) Willson-Heikkil, or
2) Zhang methods.

Of course these formats are different from ARToolKit calibration file
format. I was wondering if anybody knows how I could use the Willson-Heikkil
or Zhang formats instead of the ARToolKit calibration software. Has anybody
managed to create a robust calibration by intergrating the above? If not is
there another way that I can easily do a robust calibration?

Thanks.

Fotis Liarokapis