I'm currently working on a Java AR framework, based on ARtoolkit+ and the
JME gaming engine.
One of the features I would like to implement, is a very basic lighting
simulation. Since the borderwidth of the pattern is known, as well as the
color (black), it is possible to approach the lighting conditions in the
real environment very roughly by looking at the intesity of the pixels in
this border. For instance I could calculate the average intensity in each of
the four corners of the detected pattern, and use this information to add
lighting to the 3D scene.
I do understand that this method is far from correct, but it will produce
acceptable results in certain situations while maintaining the possibility
to render in realtime.
Anyway, my question to this mailing list is how to access the pattern image
after artoolkit has detected it. By anding this binary image result with the
original image, I would be able to extract the detected marker from the
image to start processing it. Basically I'm just hoping maybe some of you
have digged into the artk(+) code before, which would save me quite some
time and frustration :) If not, I'll digg into it myself and keep you
informed about my results.
Thanks a lot for your help.
ARToolKit mailing list