infra red - how to differentiate more than one IR beam?


    Jun 13 2006 | 12:19 am
    Hi all,
    I'm doing a project involving motion-tracking using a camera which tracks an infra-red beam. No problems there, as long as you are only tracking one thing.
    My question is: is it possible to motion track two or more different things (people, in this case) using IR at the same time? I need to be able to distinguish between the two people. In other words, it's important to know that person A is at one particular x-y coord and that person B is at another, and to follow their movements separately. Is this possible using IR?
    Hope this question makes sense. Thanks in advance for your help.
    Lori

    • Jun 13 2006 | 1:15 am
      I haven't tried this, but it just occurred to me that this might
      be a solution:
      What about building a pulse width modulation circuit on a
      small circuit board, and using it to drive an IR LED mounted on
      to it?
      With multiple boards it could be possible to set each LED to
      a different frequency, thereby allowing multiple targets to be
      discriminated using a video camera.
      IIRC, PWM is how IR remote controls work.
      What I'm not sure of though is how the frame rate of a video camera
      would interfere with this... perhaps someone more knowledgeable on
      the list would be able to clarify?
      Brian
    • Jun 13 2006 | 1:49 am
      Of course there are easier ways if the earlier suggestion is a bit
      complicated.
      I personally haven't used cv.jit yet so I can't vouch for its suitability,
      but I have successfully used an EyesWeb patch to track multiple blobs,
      sending co-ordinates as OpenSound Control messages over a network connection
      to another machine running Max.
      Although EyesWeb has good motion tracking capabilities, in assigning IDs to
      blobs it can sometimes gets 'confused' under certain circumstances in which
      case the IDs can swap around, and person A suddenly becomes person B for
      example. Jens Wunderling has written about it here:
      Just something to be aware of. Also EyesWeb is Windows only, which may or
      may
      not be a factor.
      hth,
      Brian
    • Jun 13 2006 | 2:20 am
      If you are running max on an powerPC machine and you want to track multiple
      objects, and you are not all that much into programming, you may want to
      check out softVNS2. It has objects/patches that do multiple blob tracking
      right out of the box. Check out v.heads... SoftVNS also plays nice with
      Jitter...
    • Jun 13 2006 | 7:19 am
      Using a pulse-pattern with IR will, of course, be complicated with a camera, which uses a frame-rate of 15 or 30 per second (although this wavers!). It would be possible by using very long pulse-times, but this would slow down the reaction considerably. If it is possible, it is always a good idea to use multiple sensor-data which reinforces itself. In your case, a multiple-blob tracking scheme reinforced with a much slower IR-pulse identification would work nicely. EyesWeb works well, but I usually put it on a separate machine than the one I am using for music or video projection (...still haven't tried a dual-core...) because it needs plenty of processor power.
      Dayton
    • Jun 13 2006 | 4:38 pm
      Quote: lbeckste@ryerson.ca wrote on Tue, 13 June 2006 02:19
      ----------------------------------------------------
      > Hi all,
      >
      > I'm doing a project involving motion-tracking using a camera which tracks an infra-red beam. No problems there, as long as you are only tracking one thing.
      >
      > My question is: is it possible to motion track two or more different things (people, in this case) using IR at the same time? I need to be able to distinguish between the two people. In other words, it's important to know that person A is at one particular x-y coord and that person B is at another, and to follow their movements separately. Is this possible using IR?
      How about giving one person a very powerful IR signal, and the other a "dimmer" signal so as to distinguish between the two?