Revision history [back]

I can think of two solutions:

The first is to have your 3rd node also subscribed on the camera capture and based on the binarized image coming from the 2nd node, draw on the first. This method requires synchronization to make sure that you are drawing on the correct frame, but it also depends on the accuracy you need to achieve.

The second method is to transmit the starting image through all nodes, which saves you the synchronization hassle, but adds way too much info on the network.