Calibration with calib_camera2.exe

Hi!

I'm using Windows version of ARToolkit 2.52 with VisSDK 1.2 on Win NT 4.0, 
and I have calibrated my camera using the calib_camera2.exe program.

What is the origo in the camera's coordinate system? Is it the focal point, 
the center of the lens, back plane/film plane, or other? I guess the center 
of the camera image is origo, but what about the z-direction?

After running the calib_camera2.exe, the calculated distance between the 
marker and the camera seems to be different from what I manually measure. 
How are the results of the calibration verified? Are distance-measurements 
carried out and compared to manually measured distances? In that case; how 
are the distances measured manually?

In the 2.33 alpha documentation for ARToolkit, it is presented that center 
of camera image, lens distortion and focal point is calculated.
Godding (2002) presents radial asymmetric distortion, radial symmetric 
distortion, tangential distortion among others.
I was wondering what kinds of lens distortion calib_camera2.exe finds, and 
how?

I'd appreciate any answers or links to papers on these subjects!

Best regards,
Petter

[Godding, 2002] R. Godding, Geometric Calibration and Orientation of Digital 
Imaging Systems, AICON 3D Systems GmbH, Germany, 1999.
Available from http://www.falcon.de/falcon/eng/documentation.htm 
(05.03.2002)

_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com

From:	"Petter Lundevann" <petter_lundevann@h ..........>	Received:	Mar 7, 2002
To	artoolkit@h ..................
Subject:	Calibration with calib_camera2.exe - how does it work?
Hi! I'm using Windows version of ARToolkit 2.52 with VisSDK 1.2 on Win NT 4.0, and I have calibrated my camera using the calib_camera2.exe program. What is the origo in the camera's coordinate system? Is it the focal point, the center of the lens, back plane/film plane, or other? I guess the center of the camera image is origo, but what about the z-direction? After running the calib_camera2.exe, the calculated distance between the marker and the camera seems to be different from what I manually measure. How are the results of the calibration verified? Are distance-measurements carried out and compared to manually measured distances? In that case; how are the distances measured manually? In the 2.33 alpha documentation for ARToolkit, it is presented that center of camera image, lens distortion and focal point is calculated. Godding (2002) presents radial asymmetric distortion, radial symmetric distortion, tangential distortion among others. I was wondering what kinds of lens distortion calib_camera2.exe finds, and how? I'd appreciate any answers or links to papers on these subjects! Best regards, Petter [Godding, 2002] R. Godding, Geometric Calibration and Orientation of Digital Imaging Systems, AICON 3D Systems GmbH, Germany, 1999. Available from http://www.falcon.de/falcon/eng/documentation.htm (05.03.2002) _________________________________________________________________ Join the world’s largest e-mail service with MSN Hotmail. http://www.hotmail.com

From:	Hirokazu Kato <kato@s ........................>	Received:	Mar 11, 2002
To	artoolkit@h .................., "Petter Lundevann" <petter_lundevann@h ..........>
Subject:	Re: Calibration with calib_camera2.exe - how does it work?
>What is the origo in the camera's coordinate system? Is it the focal point, >the center of the lens, back plane/film plane, or other? I guess the center >of the camera image is origo, but what about the z-direction? The origin in the camera's coordinate system is the focal point. Z-direction goes forward from the origin. >After running the calib_camera2.exe, the calculated distance between the >marker and the camera seems to be different from what I manually measure. It depends on not only the camera calibration but also marker size, marker pose and focal length of the camera. Generally ARToolKit should not be expected the accuracy of measurement result. It was designed for virtual objects overlay. If you need accurate measurement, you should use stereo measurement method. >In the 2.33 alpha documentation for ARToolkit, it is presented that center >of camera image, lens distortion and focal point is calculated. >Godding (2002) presents radial asymmetric distortion, radial symmetric >distortion, tangential distortion among others. >I was wondering what kinds of lens distortion calib_camera2.exe finds, and >how? radial symmetric distortion. ARToolKit uses dots pattern. If there are no distortion, all dots could be on the straight glid lines. So distortion parameters are adjusted so that all dots are aligned on straight lines. -- ------------------------------------------------------------------ Hirokazu Kato Faculty of Information Sciences Hiroshima City University Phone: 082-830-1705 Email: kato@s ........................ Fax: 082-830-1435 URL: http://www.sys.im.hiroshima-cu.ac.jp/people/kato/