I want to make a multi image target configuration which is not just simply a "box", but markers scattered across my room. In order to make the correct translation/positioning matrix, I figured I can use my Android and scan the room and find all projection matrices of all markers. I have to have two in view every time and it will give me the matrix of each one relative to the camera, so I know the position of the two relative to each other. That way I can find out the position of all of them and use that to get a single projection matrix.
Anyone already made such a thing before I start programming and re-invent the wheel?