Actually, I don't think this would be as difficult as you think. You
set up a global map [ie, continuous grid] of the floor involving all
of the cameras, and you simply place each blob [moving object] into a
grid cell on the map, along with its characteristics [size, color,
etc]. Then you test for consistency [in terms of charactersitics plus
position plus movement vector] as the blob moves from one grid cell to
the next with time.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.