RENCI Vis Group Multi-Touch Blog

Developing multitouch hardware and applications for research and experimentation.

Entries Comments


Duke Multi-Touch Wall Development - Hardware Part 2

22 September, 2008 (16:24) | Duke Multi-Touch Wall

Cameras

The touch detection for our direct illumination (DI) wall is based on the GC660 GigE camera from Prosilica. At VGA resolution (659 x 493), each GC660 can capture images up to 120 fps. Eight (8) of these cameras are connected via a network switch to capture the entire 13.5-foot wide x 5-foot tall screen. Each camera is fitted with Tamron 5-50mm lens and an IR filter.

Prosilica GC Series camera

Prosilica GC Series camera

IR Illumination

The IR illumination is achieved using a large number of Lorex VQ-2120 Night Vision illuminators used for IR security cameras. We also tried using the Arm Electronics IR40 IR Illuminator, but it wasn’t as bright. Dealing with hot spots on the screen was also a challenge.

Lorex QV-2120 Night Vision Illuminator

Lorex QV-2120 Night Vision Illuminator

We prototyped the wall by building the middle third with 2 projectors and 4 cameras:

Testbed Duke Multitouch Wall

Test bed for Duke Multitouch Wall

This allowed us to experiment with the cameras, illuminators and different screen materials.

Testing the wall...

Dr. Xunlei Wu testing the wall...

Projection

For the projectors, we are using the Epson PowerLite Home Cinema 1080 which gives us HD resolution at a reasonable price. With 6 projectors (3 across x 2 high) in the finished wall, we have a final resolution of 5,760 x 2,160 pixels (over 12.4 Megapixels) for a screen measuring about 13.5′ wide x 5.2′ tall.

Cameras, illuminators and a projector.

Cameras, illuminators and a projector.

Screens

Our experiments with sample screens led us towards an acrylic with a higher gain to keep the image read through the screen by the cameras as sharp as possible. We received a number of samples in acrylic as well as glass, but we decided to avoid using glass due to the risk of breakage and tempered glass didn’t transmit IR as well. Here’s our test set-up:

Screen material test set-up.

Screen material test set-up.

Screen samples.

Screen samples.

We experimented with the emulsion (projection surface) towards the user as well as towards the cameras. While we had better contrast with the emulsion towards the user, we were worried about damage. Most manufactures use a fragile sprayed-on coating for their projection surface and there was a significant risk of damage due to scratching and/or oils from the user’s hands. The screen we ended up using is 3/8″ Cinepro Ultra Diffused Screen with a 1.3 gain from RPVisuals with their Tek Satin coating on the front. In addition to reducing glare from other lights in the room, the Tek Satin coating turned out to be easier to touch - your fingers glide along it much more easily than un-coated acrylic.

On to Part 3…

« Duke Multi-touch Wall Development - Hardware Part 1

 Duke Multi-Touch Wall Development - Hardware Part 3 »

Write a comment