How exactly do the cameras work? RRS feed

  • General discussion

  • Firstly, I know that the cameras are near-infrared, and that there are five of them, but what I have not yet found good material is how they "see". Maybe this is just my missing knowledge of infrared, but when I put my hand on the table, do the cameras see the light that is reflected of my hand? If so, how on earth would sun-light affect them and NOT led lighting?

    What I'm after is a bit of reading material, if its available, because I want to start thinking about custom objects (like the monster-cup demo) that would be innovative to use, but I dont know where to start looking really.

    Thanks for any help you can give on this

    Wednesday, February 3, 2010 9:45 PM

All replies

  • Mark, I don't have any resources to point you to but I can answer a few of your questions.

    When you put your hand on the table, yes, IR light is being reflected back down into the table and picked up by the cameras. There is an IR light emitter inside the table that floods the inside top with IR light. The bright white spot that you see in the raw camera viewer in cameras 3, 4, and 5 is a reflection from the emitter.

    Sunlight is chock full of light across a very wide spectrum, most of which is not in the visible spectrum that our eyes can see. Sunlight is very bad for Surface units because it contains a LOT of IR light. LED's come in several varieties and most of the ones that are used in LED flashlights emit light in the visible spectrum only, which is why they don't affect the Surface cameras.

    A fun test is to bring up the raw camera viewer and shine a LED flashlight into the top of the unit. You won't see anything. Now, do the same thing with a 'normal' flashlight and you'll see the light in the camera view. This is because incandescent light bulbs also emit some light in the IR spectrum.

    Josh Curry (jcurry) | Sr. Support Escalation Engineer | Microsoft Surface Support
    Friday, February 5, 2010 3:19 PM
  • There's one more thing which bothers me when it comes to camera setup. As far as I know each camera covers just a piece of screen. What happens when one of them malfunctions?

    As image processing is done within the software it must have some clever algorithm of matching images from different sources (cameras). How it reacts when one of sources doesn't work?

    And would it be theoretically possible to attach more cameras (e.g. redundant ones) or the algorithm is fixed just for 5-camera setup used in Surface?

    Pawel Brodzinski

    Monday, May 24, 2010 11:18 AM