Navigateur
2010-08-03 21:26:16 UTC
I'm a beginner but I'm determined to do this.
How, in a nutshell, do I make a driver that is automatically installed
upon plugging in 2 specific webcams, that when called upon does some
real-time vision-based detection on the images (using OpenCV, my own
code) and then exposes custom-fired events and variables based on what
is found in the detection, which can be accessed system-wide by any
application, but which does not expose the images from the cameras, or
automatically install default drivers for webcams.
In other words, a vision-based user interface.
I'm aware of UVC, and AvStream.
Any ideas for ANY part of this are welcome. I'll take all ideas you
have very seriously.
Many thanks,
N
How, in a nutshell, do I make a driver that is automatically installed
upon plugging in 2 specific webcams, that when called upon does some
real-time vision-based detection on the images (using OpenCV, my own
code) and then exposes custom-fired events and variables based on what
is found in the detection, which can be accessed system-wide by any
application, but which does not expose the images from the cameras, or
automatically install default drivers for webcams.
In other words, a vision-based user interface.
I'm aware of UVC, and AvStream.
Any ideas for ANY part of this are welcome. I'll take all ideas you
have very seriously.
Many thanks,
N