...all this bumpf about how fancy they are[0] is just a load bollocks.
I am wondering if them Kinect things are really working a lot simpler; and
after waking up in the middle of a shower am now postulating that:
1. They have a simple static laser interference pattern (e.g. akin to [1]
or those
star projectors you can buy from street vendors).
2. However this one is very very fine and nicely randomish. i.e. dots less
than a few
mm appart.
3. They use a crappy low resolution normal monochrome web cam; with a
black bit of glass so
only IR gets let through.
4. They simply pass the image of this camera back.
The reason that this works is that every 'pixel' at CCD level for distances of
working range will have 1 to 100 or so 'tiny dots' on it - depending on the
distance it is at. Which is why we have roughly the range we get; why we have
such a near perfect 1/sigma callibration curve and why the range of values you
get it so odd - and why they filter certain types of noise so badly.
And perhaps, perhaps:
5. They do a phase locked loop amplifier loop in software by flashing the
projector.
But I doubt that given the noise/error artifacts.
And that is really all there is to it. Anyone here with a good high-res SRL
which can do enough IR detection to check if indeed this is the case ? I guess
a fun test would be to use a mirror to project a few extra pixels onto a flat
area - and see if that area suddenly jumps 'forward'.
Thanks,
Dw
0: http://www.wired.co.uk/magazine/archive/2010/11/features/the-game-changer
and all the mystification on how they work.
1:http://www.zimbio.com/Popular+Topics+in+Astronomy/articles/vnjstT2fTM2/Green+30mw+Laser+Pointer+Pen+Style+Star+Holographic
-
Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, please
visit http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html.
Unofficial list archive: http://www.mail-archive.com/backstage@lists.bbc.co.uk/
