As mentioned in a previous post there should be a method of getting TUIO apps to work on MS Surface. This enables apps developed for CCV based equipment to work in an MS Surface environment and gives the advantage to TUIO using developers as there is another platform they can use.

The way this should work is to make use of a part of the squidy-lib library (see squidy-lib.de) called SurfaceToTUIO to get Surface events converted into TUIO ones. However in practice although the SurfaceToTUIO C# code is available and compiles in VS2008 it does not appear to be processing Surface messages. In the Surface simulator it appears to continuously process the one event - on the surface table it does not appear to respond to events at all.

Therefore I was wondering if anyone has had experience with this and what was the outcome. I hope to track the errors down and post the results on this forum.

Although strictly speaking this should go on the squidy-lib forum there has not been a single read of my post on there so this certainly seems to be the more active community!

bamalam

First off, my testing environment. Running Ubuntu 8.10 AMD64 with a custom kernel. Note: the standard issue model should do just fine, mine is trimmed down and optimized for my processor. These instructions should apply just fine to other distributions, but you will have to modify the commands accordingly (such as using root instead of sudo, and your package manager). I am going to assume that you have a basic knowledge of the terminal and all necessary build tools installed ("sudo apt-get install build-essential” on Ubuntu).

Now, on to the fun stuff. Thanks to a group on Linuxtv.org, we have a nice new V4L driver available for this. The first step is to install ‘hg’ (included in the “mercurial” package on Ubuntu), and download the latest version of said driver:

You should have a folder named “gspca” in the current terminal directory. Now we want to change into that directory and begin the build process:

At this point, make might complain about any number of things. You will want to check and make sure it is installing for your *current kernel* (check with “uname -r” if you are unsure). That is the biggest issue. If it complains about something else, post here and I or someone else will try to help as much as possible. If your build is successful, go ahead and install it:

You should now have a patched V4L installed, and the driver should be working. Let’s go ahead and test this. Change into the test programs directory (assuming you are still in “gspca"), and build the driver test file:

Now plug in your camera, and it is time to see if the driver install worked.

. This should produce an output string with information about the file format, and what is being transferred, etc.. If it says “no /dev/video*”, make sure the camera is plugged in. It seems to autodetect the proper video device, so other cameras or video capture cards shouldn’t be a problem.

If you have made it this far, congratulations. Time to see some actual results in video form. Install one of the configurable video programs (Mplayer, VLC, Xawtv, etc..). VLC is probably the easiest to use here so:

Open VLC, then go to Media > Open Capture Device. In the configuration window, set it to V4L2, video device: /dev/video0, audio device: /dev/audio1. Your video and audio devices may be different, this assumes no other video inputs and one audio input (motherboard mic input in my case). In a second or two, you should see some nice 640x480 video in proper color, and the audio should be working. This concludes this part of the install process, as I haven’t yet tried anything further myself (will test tbeta and some other programs soon and report back).

The problems seem to be: jumpy video, and VLC crashes if custom settings are used (e.g. for resolution and framerate). The occasional jumpy video can be blamed on the huge amount of raw video data being processed, and I am still looking into the custom settings problem. I imagine that as drivers for this camera progress, performance will too.

This concludes my tutorial, I hope it helps some. Please feel free to chime in with any questions or comments.

[edit] Oops, almost forgot to give credit. I had nothing to do with developing drivers, the guys at linuxtv.org (specifically “jfancois”, but I am sure there are others) took care of that. Getting it to work was largely thanks to excerpts from the PS2 Dev forums, specifically “jimparis” and “ao2”. Thread: http://forums.ps2dev.org/viewtopic.php?t=9238&postdays=0&postorder=asc&start=0

But I’m not writing about Magnetic, I’m writing to ask what is the best approach to remap the camera image to the image you analyze for blobs?  CCV seems to work well with both linear and lens distorted images (so long as there are more points), whereas mine sucks with a wide angle lens.  Does CCV use any undistorting techniques, or are y’all just using cvFindHomography?  I’m doing the latter, and just wanted to figure out which route to pursue.

My options are either:

A) Continue to use cvFindHomography() and allows users to add more points (a la CCV).
B) Stick with just four corners and then use arrow keys to adjust radial distortion values.

I’m thinking A would be the right route, as B could fail if the image is lopsided.  It’s important that works for any type of screen.  In fact, half of Magnetic’s whole purpose is to track limbs, etc, not touch points, so dedicated radial distortion correction is not an option.

Ideas?

A few early Magnetic videos, if interested:
http://www.youtube.com/watch?v=AEysenAU3qU
http://www.youtube.com/watch?v=W8kR7jbbtpY
http://www.youtube.com/watch?v=9NUa3k6cWY4

I’ve been lurking on this site for weeks now, and any search query i think of Google will lead me here.. So this is probably the forum that knows Everything about anything.

I decided to create this account so I can participate in discussions and fire my questions when I actually done my research and started putting A and B together :)

I am a grad student on the Utrecht School of the Arts (http://www.hku.nl) and studying Digital Media Design.

Hopefully you’ll see me around.

greetings Marnix

There are great solutions for blob tracking in the machine vision industry that can fit perfectly in a MT project.
Many of them support multicameras, amazing frame rates and very low latency.
We are looking for a solution of this kind for a large rear DI.

Just a couple of example :

http://www.dalsa.com/ipd/products/boa.aspx
http://www.analogic-computers.com/ProdServ/Bi-i/Bi-iv301F.html

I am trying to build a LCD DSI setup.

I have bought Endlighten plexi and led strips (850nm )which I mounted on the edges. I put a camera under, and a glass over, but i cant see the DSI effect. If i press the finger directly on the plexi i can see the FTIR effect. The edges of the plexi are milky so i think there are endlighten. I testes this with light leds and the light is filling the whole plate. What am I doing wrong? I cant i see the fingers on the touch surface?

Please help!!!!

I wander for quite a while on the internet and I’ve found these setup which are quite quick to be put in place :
PlayAnywhere of Microsoft Research : http://www.youtube.com/watch?v=0OomjJpa5fk
LightTouch of light blue optics : http://lightblueoptics.com/
And of course the concept of Hype : http://nuigroup.com/log/comments/portable_large_interactive_display/

So here is my question, Is anyone know what software they are using? More particularly is anyone have made something similar in the technology as the PlayingAnywhere projet?

Thanks

PS: sorry for my english, i’m french!!!

I’ve got a standard 13.8V - 20 A power supplied.
I want to try to test my webcam, so I made a little circuit of 8 leds SFH845P and a 18 Ohm resistor (thanks to ledcalculator.net)

So my webcam is working great ! But my leds are very bright at the beginning and then the intensity of the infrareds are getting down. I even killed the last led before the resistor.
I have very little knowledge as you can see because I have probably made a huge mistake but can figure out what ... before wiring everything I want to be sure of what I am doing .

I plan to build a circuit with 84 leds, results on led calculator seems clear for me but maybe I should use another power supply.

Many thanks,
Cédric