Multicamera ob ject disambiguation using view angles in a plane

Abstract : Suppose three cameras able to identify the angles of view to the same n objects in the plane of the cameras but unable to identify the objects. The problem is to put the angles into correspondence, ie. find n points in the plane each covering one line of view from each camera and each line of view covered. We compare the performance of two algorithms, one for the static case and the other one, possibly distributed on the cameras, for moving objects.