Dear all,

regarding ARMulti: I am having the same questions as Enrico, mainly this
one:
> what does the matrix relative to each marker (in marker.dat) describe?

I have markers on walls, which should serve as a background to calculate the
camera position in a global coordination system.
Similar to the shared white board described in
www.sys.im.hiroshima-cu.ac.jp/people/kato/publications/c-99.IWAR99.pdf , but
on three sides.

Just one simple example of how to use the matrix would be of great help.

#marker 1
Data/multi/patt.a
40.0
0.0 0.0
1.0000 0.0000 0.0000 0.0000
0.0000 1.0000 0.0000 0.0000
0.0000 0.0000 1.0000 0.0000

Thanks - Maia

Hi,

the simple example is already given in the ARToolKit example. Ok, lets
have a more detailed look at the marker.dat file. It describes a set of
markers plus a offset transformation for each marker. This offset
transformation is applied to the transformation matrix of the detected
marker.

The result is a referenze coordinate system for all these markers. This 
means if we detect one of the defined markers then we always get this 
coordinate system.

If we want to define a new multi marker then we have to define this 
reference coordinate system. In marker.dat the marker Data/multi/patt.a 
is such a reference coordinate system, because the transformation is the 
identity matrix. If we take a closer look at the other markers we will 
see that:

* Data/multi/patt.b is translated 100 along the x-axis
* Data/multi/patt.c is translated 200 along the x-axis
* Data/multi/patt.d is translated -100 along the y-axis
* Data/multi/patt.g is translated -100 along the y-axis
   and 100 along the x-axis
* Data/multi/patt.f is translated -100 along the y-axis
   and 200 along the x-axis

So, all these markers are placed by the given offset transformations.
For the MR assembly instructor (presented at ISMAR03) we use a similar
approach and we automatically calibrate these transformations. Just by
detecting one of the already defined markers and calculating the
tranformation between the reference system and the new marker.

I hope my explanation helps a little bit.

Best regards,
Juergen Zauner

============================================
Upper Austria University of Applied Sciences
Department of Mediatechnologies and -design

Hauptstraße 117
A-4232 Hagenberg
AUSTRIA

Hi,

the simple example is already given in the ARToolKit example. Ok, lets
have a more detailed look at the marker.dat file. It describes a set of
markers plus a offset transformation for each marker. This offset
transformation is applied to the transformation matrix of the detected
marker.

The result is a referenze coordinate system for all these markers. This
means if we detect one of the defined markers then we always get this
coordinate system.

If we want to define a new multi marker then we have to define this
reference coordinate system. In marker.dat the marker Data/multi/patt.a
is such a reference coordinate system, because the transformation is the
identity matrix. If we take a closer look at the other markers we will
see that:

* Data/multi/patt.b is translated 100 along the x-axis
* Data/multi/patt.c is translated 200 along the x-axis
* Data/multi/patt.d is translated -100 along the y-axis
* Data/multi/patt.g is translated -100 along the y-axis
   and 100 along the x-axis
* Data/multi/patt.f is translated -100 along the y-axis
   and 200 along the x-axis

So, all these markers are placed by the given offset transformations.
For the MR assembly instructor (presented at ISMAR03) we use a similar
approach and we automatically calibrate these transformations. Just by
detecting one of the already defined markers and calculating the
tranformation between the reference system and the new marker.

I hope my explanation helps a little bit.

Best regards,
Juergen Zauner

============================================
Upper Austria University of Applied Sciences
Department of Mediatechnologies and -design

Hauptstraße 117
A-4232 Hagenberg
AUSTRIA