Analyse video signal


we have this: an webcam, which is in front of a white wall. An dark oject is between the webcam and the wall. From this object we need (only) the coordinates of the top edges. This edge-coordinates (where background meets foreground) should be in a matrix (or whatever) to make an collision control.
At the moment we have the videoIn-signal, but we despair of getting coordinates out from this… please give us a hint how to go on!


contour and colortracker should both be able to do that

Thanks, with contour it worked out perfect. We get all coordinates from the outline and we realized thats what we need.