Brian Lucid

  • Home
  • About
  • Departments
    • Actionscript
    • Multitouch
  • Subscribe via RSS

ReacTIVision 1.4 quick test

March 1st, 2010  |  Published in Multitouch

Most of the blob-tracking tests I have done with the table have been done with CCV. One of the original goals of the table, however, was also to support fiducial tracking. Above is a screen-shot of ReacTIVision identifying a page of fiducials. This was a very quick test, laying a single page of fiducials on top of my projection layer. I did no calibration of ReacTIVision whatsoever. ReacTIVision ran at 31 frames per second.

ReacTIVision was not picking up finger-blobs, but I figure that can be corrected with proper camera calibration.

Projection Test #2

January 17th, 2010  |  Published in Multitouch

WIth the aluminum frame in place, touches register much more strongly. Blobs are so strong that they are getting tracked before the finger comes in contact with the surface (hovering), so this will mean some re-calibration.

The table was set up today at 33-inches in height which gives me a 41-inch projection area on my 50-inch surface.

I spent an hour with a bounce mirror trying to fold the projector beam into a smaller space, but the right combination is proving elusive. I am going to make a paper prototype and do some test folding to nail down some of the angles…

U-channel

January 11th, 2010  |  Published in Multitouch

I have been looking at various sites on the internet for aluminum U-channel with a .5-inch opening. This slided over the ends of my endlighten and projection layer, blocking the IR LEDs from the camera. Unfortunately, the sources I found were somewhat expensive, and shipping costs were prohibitive.

Therefore, I was pleasantly surprised to find inexpensive ($9.39) 96-inch U-channel (with a .5-inch opening) at the local Home Depot.

Projector Calibration Test

January 7th, 2010  |  Published in Multitouch

Performed a quick test bringing together the projector with CCV blob tracking. I have not done any testing with projector positioning and the bounce-mirror, so I simply pointed the projector vertically up at the screen. The projector gives off little to no IR radiation, so there were no hot spots caused by the projection, which is very good news.

I did not have too much trouble calibrating CCV to align the touches with the screen. As the IR strip is not well affixed to the EndLighten, I did get a bit of extra IR light on some areas of the surface creating false blobs.

Once calibrated, finger tracking worked quite well. CCV is currently only running at 30FPS on my older Macintosh MacBook Pro. The i7 PC runs CCV at 60FPS so I should see a tracking improvement once I get the PC inside the table.

First blob tracking test

January 6th, 2010  |  Published in Multitouch

The above image shows the illuminated EndLighten with the rear-projection acrylic placed on top of it as a touch surface. As you can see, the projection layer makes the blobs quite a bit softer.

A screenshot of CCV showing the touch-blobs being tracked. Click the image to see my current settings. I found it to be much easier to track blobs with the camera at its full 640 x 480 resolution.

Illuminating the EndLighten Acrylic

January 6th, 2010  |  Published in Multitouch

For a rough test, I quickly wrapped the IR strip around the edge of the EndLighten sheet and held it in place loosely with tape. As you can see, there is a lot of light lost here. I will be using u-channel to better position and align the strip lighting to the acrylic.

Shot with the IR camera, you can see how the sheet illuminates evenly when lit on edge.

Here you can see the white blobs generated when a finger comes in contact with the illuminated surface. The camera is currently sitting 29 inches away from the surface using a 2.8mm lens.

Environmental Lighting IR Strip Test

January 6th, 2010  |  Published in Multitouch

Above is an IR cam capture of my first test of the 850nm IR LED strip from Environmental Lights.

The LEDs are mostly invisible to the human eye when illuminated. When shot with a digital camera, each LED gives off a tiny bit of red light.

Acrylic Edge Polishing

January 6th, 2010  |  Published in Multitouch

Rough-cut acrylic does not have the smooth edge required for good IR illumination / penetration of the EndLighten. So, I needed to get the edges of the plexiglass polished.

While this can be done by hand, I opted to have mine done professionally. After a bit of research I found Altec Plastics (800 477 8196), located inside a beautiful old fire station in South Boston. My 43 x 27 inch sheet cost $20 to polish, and they were able to turn the project around in an afternoon.

Parts: Toshiba EW25U DLP projector

December 23rd, 2009  |  Published in Multitouch

The table will be built around a 2,600 lumen Toshiba TDP EW25U DLP projector. This WXGA (1280 x 800 16:10) projector is classified as an “ultra-short-throw” projector, needing only two feet to project a 50″ image.

Recently discontinued, the projector can still be found online. We bought ours from Projector People

Projector People has a useful projection calculator for this projector, available from the “Complete Specs” product page.

Parts: 850nm LED Light Strip

December 23rd, 2009  |  Published in Multitouch

Instead of manually wiring together IR LEDs, the table will use an Infra Red (850 nm) LED light strip from Environmental Lights for its IR light source. This strip will be wrapped around the edges of the EndLighten acrylic.

Parts: Modified PS3 Cam

December 23rd, 2009  |  Published in Multitouch

To track finger touches the table will use a PS3 Eye webcam modified by Peau Productions.

The PlayStation Eye is an relatively inexpensive webcam from Sony capable of capturing 60 frames per second at a 640×480 pixel resolution, and 120 frames per second at 320×240 pixels. For the purposes of multi-touch blob tracking, fast frame rates provide a better response.

This camera has had its IR filter removed and had a visible-light filter put in its place so that it only captures light in the 850nm Infrared range. It has also been given an M12 mount so that lenses of different focal lengths to be used.

Parts: ACRYLITE Rear Projection Sheet: 7D513 RP

December 22nd, 2009  |  Published in Multitouch

The touch surface of the table will lie on top of the EndLighten. This sheet also serves as the “projection layer”, i.e. the frosted material that will trap and hold the image from the projector.

I am using an acrylic sheet specially designed for rear-projection: ACRYLITE® Rear Projection Grey 7D513 RP. As an added benefit, it has a nice texture.

The sheet is 0.118 of an inch thick. Sized at 43″ x 27″. It cost about $182.00 USD from the Evonik online store.

Parts: Acrylite EndLighten XL

December 21st, 2009  |  Published in Multitouch

Here is our sheet of Acrylite EndLighten XL, wrapped in its protective coverings. It is sized at 43″ by 27″ based upon the 16:10 aspect ratio of our chosen short-throw projector. While optically clear, the surface of this material becomes infused with light when lit from the edges of the material due to being embedded with tiny particles that act like mirrors.

The EndLighten was purchased from Evonik via their online store. The sheet cost about $206.00 USD.

The acrylic is delivered with a cut edge from the Evonik store. Next step is to polish this up to an optical finish to best allow the IR lighting to penetrate the surface.

Parts: i7 Gaming PC

December 21st, 2009  |  Published in Multitouch

The PC above will be responsible for both running the blob-tracking software (CCV) and the actual multi-touch application (which students will write in Flash/AS3 or Processing)

Project: A 53″ DSI Multi-touch surface

December 20th, 2009  |  Published in Multitouch

I am currently building a large-scale optical multi-touch table for the Dynamic Media Institute at the Massachusetts College of Art and Design. The goal is to have this table up and running early in the semester so it can be integrated into our upcoming Graduate Studio 2 course.

In it’s simplest form, an “optical multi-touch system” consists of an infrared camera or optical sensor, infrared light sources, and some sort of presentation technology such as a standard video projector or LCD screen.

Infrared cameras and light sources are used to illuminate and capture finger touches and movements upon a surface. This is done in the wavelengths beyond the range of the human eye so that it does not interfere with the visual image on the touch surface.

The table being built for use at the Dynamic Media Institute uses an optical technique called Diffused Surface Illumination (DSI). This technique employs a special type of acrylic — named EndLighten — that is embedded with tiny particles that act like mirrors. While optically clear, the surface of this material becomes infused with light when lit from the edges of the material.

In the DMI table the acrylic will be illuminated with IR light along each edge, resulting in an even plane of IR light across the surface of the table. When touched, finger presses will appear as “blobs” that can be recognized via the IR camera and sent to software coded to track blobs running on the PC inside the table.

More information about optical multi-touch techniques is available in the “Community Book” authored and hosted by the community at NUIgroup.com.

Building your first multi-touch application…

December 18th, 2009  |  Published in Actionscript, Multitouch

This tutorial serves as an excellent starting point to using the As3 multi-touch libraries available at nuigroup.com.

Previously


Jan 17, 2010
Projection Test #2

by Brian Lucid | Read | No Comments

WIth the aluminum frame in place, touches register much more strongly. Blobs are so strong that they are getting tracked before the finger comes in contact with the surface (hovering), so this will mean some re-calibration.
The table was set up today at 33-inches in height which gives me a 41-inch projection area on my [...]


Jan 11, 2010
U-channel

by Brian Lucid | Read | No Comments

I have been looking at various sites on the internet for aluminum U-channel with a .5-inch opening. This slided over the ends of my endlighten and projection layer, blocking the IR LEDs from the camera. Unfortunately, the sources I found were somewhat expensive, and shipping costs were prohibitive.
Therefore, I was pleasantly surprised to find [...]


Jan 7, 2010
Projector Calibration Test

by Brian Lucid | Read | No Comments

Performed a quick test bringing together the projector with CCV blob tracking. I have not done any testing with projector positioning and the bounce-mirror, so I simply pointed the projector vertically up at the screen. The projector gives off little to no IR radiation, so there were no hot spots caused by the projection, which [...]


Jan 6, 2010
First blob tracking test

by Brian Lucid | Read | No Comments

The above image shows the illuminated EndLighten with the rear-projection acrylic placed on top of it as a touch surface. As you can see, the projection layer makes the blobs quite a bit softer.

A screenshot of CCV showing the touch-blobs being tracked. Click the image to see my current settings. I found it to be [...]


Jan 6, 2010
Illuminating the EndLighten Acrylic

by Brian Lucid | Read | No Comments

For a rough test, I quickly wrapped the IR strip around the edge of the EndLighten sheet and held it in place loosely with tape. As you can see, there is a lot of light lost here. I will be using u-channel to better position and align the strip lighting to the acrylic.

Shot with [...]


Jan 6, 2010
Environmental Lighting IR Strip Test

by Brian Lucid | Read | No Comments

Above is an IR cam capture of my first test of the 850nm IR LED strip from Environmental Lights.
The LEDs are mostly invisible to the human eye when illuminated. When shot with a digital camera, each LED gives off a tiny bit of red light.

About Brian Lucid

Brian Lucid is a designer, educator, and consultant whose personal and professional work spans from traditional communications in static and temporal media to computational design and physical/digital interfaces.

Categories

  • Actionscript
  • Multitouch

Contributors

  • Brian Lucid

Popular

  • Blogroll

    • As3 for Multitouch
    • ActionScript.org forums
    • Aixiz (Laser Suppliers)
    • Dynamic Media Institute
    • Massachusetts College of Art and Design
    • Peau Productions
    • Tom Igoe's Code Blog


  • ©2010 Brian Lucid
    Powered by WordPress using the Gridline Lite theme by Graph Paper Press.