Duke Multi-Touch Wall Development - System
The Duke Multi-Touch Wall uses 8 cameras to detect touches across the large 13.5-foot wide x 5-foot tall screen. These cameras are connected via network cables to a Camera Control Node where a separate instance of Touchlib is running for each camera. The Camera Control Node handles the image processing and blob tracking for each camera separately. These touch events are all then routed to another “composition” process which eliminates duplicate touches due to overlapping camera boundaries and assigns unique touch IDs.
This process then sends the stream of unique touches to the gesture engine (running on the Windows or Linux display machine) which interpret them as gesture events. These gesture events are finally sent to the client application to manipulate windows, data, etc.