Stage Five
I hooked the system up to a projector to try and get some sense of scale using the graphics card on my desktop and it worked really smoothly. I still couldn't get the camera far enough off the floor to allow for a wide angle. But as a prototype this is pretty good:

Stage Four
I have gone back through the MaxMSP structure and made the selection of specific colours more accurate and it seems to have done the job, especially with the selection of the green. Take a look:

Stage Three
I set the project up in college on Monday morning with a projector, and although it all worked...it wasn't as smooth as I was hoping. The key issues were:

a) The view of the camera was amazingly small, it was at least 8-9 feet off the floor and yet it could only see about 1.3 metres of the floor. This could just be the quality of the camera, but could easily be over come I'm sure with a lens or a better camera.

b) Because it was in college, I had to run the thing through my laptop, which despite it chipper integrity...isn't really up for this sort of processing, so it was incredibly slow.

Stage Two
At this stage I linked the coloured movieclips with some php to pull news from an RSS feed and photos from a folder, although if system speed weren't an issue this could also be done from an RSS feed like Picasa or Flickr etc.

Stage One
This first video was taken once I got Flash to recognise the position of the lights from the database, and place the correctly coloured movieclips accordingly.