> Anyway, I believe those projectors, like TVs and monitors, actually
> perform a sweep over the screen, and we see it as a whole image due
> to... guess what, persistence of vision ;-) (and maybe to screen
> persistence on TVs and monitors). What would probably be needed is to
> slow down the scanning in the projector or to use a very high speed
> camera.

No, those projectors use LCD panels inside, every pixel is on continuously,
and a whole line is updated at a time.  So if you like it "sweeps"
downwards, but you will never see any flicker because the pixels don't "turn
off" like they do in a CRT.

Post a reply to this message

Jim Henderson wrote:
> On Tue, 10 May 2005 22:18:13 +0100, scott wrote:
>
>> I don't see why it isn't possible.  The projector is lighting one
>> tiny segment of the card at a time, if it hits a red bit, the book
>> is going to be lit with a red hue, very easy to pick up by the
>> camera.  Once it's scanned the whole card, it would be easy to build
>> up the picture.
>
> That's the trick, though - the projector shown in the video, unless
> I'm terribly mistaken, is just a standard video projector (the type
> you might plug a PC into to project it on the screen) - so the
> projector is in fact lighting the entire book/card scene all at the
> same time.

You are not mistaken, it is a *projector* and not a light bulb exactly so
they can light up one point at a time.  They just display a black image on
the projector with one white pixel, and move that pixel about.  It's the
same as having a very narrow spotlight and moving it about.

Post a reply to this message

Ah, yes, that does make a difference - after reading the paper, this makes
perfect sense. :-)

Jim

Post a reply to this message

"Jim Henderson" <nos### [at] nospamcom> wrote in message
news:pan### [at] nospamcom
> Ah, yes, that does make a difference - after reading the paper,
> this makes perfect sense. :-)

Projectors are often used to capture 3D data using a normal camera.  If you 
place the projector and camera at slightly different positions, you can then 
project special patterns using the projector.  For example, if you project a 
series of images using the projector:

11110000      11001100     10101010
11110000 -->  11001100 --> 10101010  (1=white)
11110000      11001100     10101010  (0=black)

Then record the three images the camera sees in monochrome.  For each pixel 
in the camera, it will have three values (one from each frame), this is 
effectively a 3-bit number that will tell you which line on the projector is 
lighting that pixel.  Using some simple geometry you can then work out the 
distance of that point from the camera (because you know the position of the 
camera and the projector, and the angle that the light came out / went into 
each).

Do this for every pixel and you have a height-field of your scene built up 
very quickly.  If you projector has 1024 vertical lines, you only need 
project and record 10 frames to get the height-field, that's less than 1 
second.

Do the whole process at several different known locations and you start to 
build up a very detailed 3D map very quickly...

Projectors are almost cheap enough now for me to not feel guilty buying one 
for home use :-)

Post a reply to this message

Among other things, Jellby saw fit to write:

> Here is a physical technique to interchange the roles of a camera and a
> light-source which is reminds me of how POV-Ray calculates the rays.

I've tried implementing this in POV-Ray, see p.b.images: "Dual photography
concept testing".

-- 
light_source{9+9*x,1}camera{orthographic look_at(1-y)/4angle 30location
9/4-z*4}light_source{-9*z,1}union{box{.9-z.1+x clipped_by{plane{2+y-4*x
0}}}box{z-y-.1.1+z}box{-.1.1+x}box{.1z-.1}pigment{rgb<.8.2,1>}}//Jellby

Post a reply to this message