In our experience, the ARToolkit (as released) supports a single 
camera, and doesn't behave well if you try running multiple video 
streams through it.  The stereo ARTk programs/demos I saw used two 
separate machines.

Our thought was to run two instances ARTk in two subprocesses (threads) 
with their own data forks, and  communicate via shared memory.  We 
never tried it though.


On Monday, May 12, 2003, at 11:43 AM, Fivos DOGANIS wrote:

> Hello,
>
> Has anyone tried the stereo functions of the ARToolKit?
> I have been trying to integrate ARToolKit into a non glut application 
> and
> everything was working fine as long as I used only one camera.
>
> Now I would like to use two cameras and it seems that the 
> arsInitCparam and
> arsDetectMarker functions don't work as expected.
>
> Dr. Kato, are the stereo functions finished and tested or not? I 
> remember
> you were saying that this was 'future work'.
>
> Thanks for any replies!
>
> Regards,
>
> Fivos

Hello,

Has anyone tried the stereo functions of the ARToolKit?
I have been trying to integrate ARToolKit into a non glut application and
everything was working fine as long as I used only one camera.

Now I would like to use two cameras and it seems that the arsInitCparam and
arsDetectMarker functions don't work as expected.

Dr. Kato, are the stereo functions finished and tested or not? I remember
you were saying that this was 'future work'.

Thanks for any replies!

Regards,

Fivos

Thanks for your fast reply!

I have tried indeed to run ARToolKit functions multiple times and my program
crashed. After looking at the source code, I noticed some 'nasty' extern
variables :)

I therefore tried to use the ARSParam variable instead of the ARParam but
could not display anything on screen... Although I understand what each
extern variable is, I don't know when thay are modified etc.

I agree taht the best thing would be to use the threads you mentionned, and
use one ARToolKit instance per thread! I might have a look again at
OpenTracker and other documents I overlooked in the past!

Thanks again!

Fivos

-----Message d'origine-----
De : Blair MacIntyre [mailto:blair@c ............]
Envoye : lundi 12 mai 2003 18:04
A : Fivos DOGANIS
Cc : artoolkit@h ..................
Objet : Re: ARToolKit in Stereo


In our experience, the ARToolkit (as released) supports a single
camera, and doesn't behave well if you try running multiple video
streams through it.  The stereo ARTk programs/demos I saw used two
separate machines.

Our thought was to run two instances ARTk in two subprocesses (threads)
with their own data forks, and  communicate via shared memory.  We
never tried it though.


On Monday, May 12, 2003, at 11:43 AM, Fivos DOGANIS wrote:

> Hello,
>
> Has anyone tried the stereo functions of the ARToolKit?
> I have been trying to integrate ARToolKit into a non glut application
> and
> everything was working fine as long as I used only one camera.
>
> Now I would like to use two cameras and it seems that the
> arsInitCparam and
> arsDetectMarker functions don't work as expected.
>
> Dr. Kato, are the stereo functions finished and tested or not? I
> remember
> you were saying that this was 'future work'.
>
> Thanks for any replies!
>
> Regards,
>
> Fivos

Dear Dr. Kato,

Thank you very much for your source code and your fast reply! What I really
needed was an example of how to use ARToolKit in stereo. Therefore your
'stereo.c' example was very useful! Now I can at last track patterns with
two synchronised cameras under WIN32! Actually, I use the ARToolKit for
tracking only (i.e. find the pixel coordinates of the patterns) and I have
my own video and camera calibration routines.

> Now I prepared a calibration utility 'calib_stereo' which needs very
careful
> manual operations like 'calib_cparam'.

Wouldn't it be great to replace this delicate procedure with something like
OpenCV or Zhang's calibration method? I get really good results using those
without much difficulty. Anyway, the ARToolKit works fine with uncalibrated
cameras as well (for my purposes at least!).

Thanks again!

Fivos DOGANIS

-----Message d'origine-----
De : Hirokazu Kato [mailto:kato@s ...................]
Envoye : mardi 13 mai 2003 07:06
A : Fivos DOGANIS
Cc : artoolkit@h ..................
Objet : Re: ARToolKit in Stereo


Hi,

>Dr. Kato, are the stereo functions finished and tested or not? I remember
>you were saying that this was 'future work'.

Yes. I am developing a stereo version of ARToolKit.
Basically I finished to make the functions for stereo tracking.
But I have not finished to build some utilities and documents
about calibration for stereo cameras yet. It is required to calibrate
the relationship between two cameras as well as intrinsic parameters.
Now I prepared a calibration utility 'calib_stereo' which needs very careful
manual operations like 'calib_cparam'.
If you are interested in it in spite of these difficulties,
you can download it from
http://www.sys.im.hiroshima-cu.ac.jp/people/kato/etc/ARToolKit3.271.tar.gz
It can work on Linux with two analog NTSC cameras.

--
------------------------------------------------------------------
  Hirokazu Kato
  Graduate School of Engineering Science
  Osaka University                   Phone: +81-6-6850-6381
  Email: kato@s ...................     Fax: +81-6-6850-6341
  URL: http://www.sys.im.hiroshima-cu.ac.jp/people/kato/

I promise I'll give it a shot after I finish my app... Camera calibration is
usually very boring and time consuming :)

Fivos

-----Message d'origine-----
De : Hirokazu Kato [mailto:kato@s ...................]
Envoye : mercredi 14 mai 2003 02:41
A : Fivos DOGANIS
Cc : artoolkit@h ..................
Objet : RE: ARToolKit in Stereo


>Wouldn't it be great to replace this delicate procedure with something like
>OpenCV or Zhang's calibration method?

I think it is very good idea. But I do not have time to do.
I hope someone will do that and put it on the web.

--
------------------------------------------------------------------
  Hirokazu Kato
  Graduate School of Engineering Science
  Osaka University                   Phone: +81-6-6850-6381
  Email: kato@s ...................     Fax: +81-6-6850-6341
  URL: http://www.sys.im.hiroshima-cu.ac.jp/people/kato/