0

A video-camera is mounted opposite a blackboard on which several concentric circles are roughly drawn in chalk. The circles are non-ideal, and the camera is not quite steady. The camera has a CVBS (composite) output.

I need to make a mission-specific hardware that would take snapshots from the camera, calculate circle diameters, and output them to a simple LED display.

So I have to: 1) capture an image; 2) apply a few alogorithms to improve the image (like thinning etc.); 3) recognize & measure the circles.

Since I'm totally new to this area, I would appreciate any inputs: schematic description of how such a circuit can be designed, keywords to google-around, examples, existing solutions that can be re-used -- I'd appreciate any pointer.

P.S. No, it's not a homework :). Just thinking of an idea and trying to estimate its complexity.

Igor R.
  • 14,716
  • 2
  • 49
  • 83
  • 1
    Why not prototype this on a suitable device first, e.g. Android phone or iPhone ? No need to design any hardware. – Paul R Oct 24 '12 at 19:03
  • 1
    I would recommend using matlab or opencv (would run on android and iPhone like Paul suggests) to prototype your algorithm, then worry about hardware once you know exactly what you need to be able to recognize your circles. – Hammer Oct 24 '12 at 19:08
  • @Hammer yes, of course the whole process should be simulated first on PC or any suitable device. But the main question is whether it's eventually feasible to create such a hardware, and how to. – Igor R. Oct 24 '12 at 19:42
  • I see, that is well beyond my expertise. You might have better luck with a question like that on the signal processing stack exchange site – Hammer Oct 24 '12 at 19:53
  • 1
    You might want to look at e.g. Arduino with a suitable camera shield. – Paul R Oct 25 '12 at 05:44
  • @Paul R great, thanks for the pointer! – Igor R. Oct 25 '12 at 09:26

0 Answers0