What I would like to obtain is the position (X, Y and maybe Z, if it’s possible) of three little objects in a small room (I’m thinking about three spheres with a diameter of 10-20 cm).

But I would like that there will not be any cable between these objects and my laptop, so I was thinking about some infrared sensor.

I already own an Eroktronix Miditron, that maybe could be useful.

Can somebody give me some tips please?

Cheers

I agree with the IR cam idea, unless you need positions more often than 5 times a second. Also, you would need multiple cameras to get 3D

Unibrain makes a Fire-i camera that doesn’t have an IR filter:

http://www.unibrain.com/Products/VisionImg/Fire_i_DC.htm

You can make a visible light filter with exposed color film. (Lots of info on google)

cv.jit is handy for tracking:

http://www.iamas.ac.jp/~jovan02/cv/

If you don’t want IR LED’s on the objects you are tracking, you can put some reflective material on them, and bathe the whole scene in IR.

I haven’t used it, but this system looks interesting:

http://www.naturalpoint.com/optitrack/

mz

Michael thanks for your links, however IR cam seems to be an expensive and complex solution (349$ for Optitrack with only one camera, without 3D).

Or something like Makethings Infrared Distance Measuring Sensor ( http://www.makingthings.com/products/ACC-IR-DIST-001 ) works only up to 150 cm.

I’m thinking I could use some videotracking system (jit.findbounds or cv.jit.blobs.centroids or something else).
So, I would use a firewire camera in order to track any stained and shining object inside the dark room.

Maybe I would put an electric torch inside a semitrasparent stained objects (a green, a red and a blue one). I could paint three plexiglass spheres, if it will be possible to re-open them in order to change batteries of torches.

So I would haven’t galling cables inside the room. But as compensation I would have to change batteries every two hours.

Any further suggestions will be appreciated.

Quote: mzed wrote on Wed, 18 April 2007 19:45
—————————————————-
> I agree with the IR cam idea, unless you need positions more often than 5 times a second. Also, you would need multiple cameras to get 3D
>
> Unibrain makes a Fire-i camera that doesn’t have an IR filter:
> http://www.unibrain.com/Products/VisionImg/Fire_i_DC.htm
>
> You can make a visible light filter with exposed color film. (Lots of info on google)
>
> cv.jit is handy for tracking:
> http://www.iamas.ac.jp/~jovan02/cv/
>
> If you don’t want IR LED’s on the objects you are tracking, you can put some reflective material on them, and bathe the whole scene in IR.
>
> I haven’t used it, but this system looks interesting:
>
> http://www.naturalpoint.com/optitrack/
>
> mz
>
—————————————————-

You should use LED’s and change the batteries daily instead.

Thank you… so do you think that a 40-50 lm led would last one day long?

http://infusionsystems.com/catalog/product_info.php/cPath/24/products_id/159?osCsid=106fb31342173843921c8c24f4d72863

or InfusionSystems MoveOn:

http://infusionsystems.com/catalog/product_info.php/cPath/24/products_id/38?osCsid=106fb31342173843921c8c24f4d72863

The problem with IR sensors, I guess, will be that if a person stays between the IR-cam and the reflective sphere, the IR-cam will be not able to recognize it. Isn’t it?

Does anybody try Infusion System sensors and can give me some tip?

Thank you again

Gerald