Blob tracking and collision detection

All of the examples I’ve come across, including those in your link say all you need to know is the x,y , width and height of your object(s). My biggest confusion is then how do I find these values when all the objects I’m looking at exist in the same endless array.