forums.naturalpoint.com

Hey,

I recently got a SmartNAV 4 camera and have been experimenting with the OptiTrack SDK, to try and see if Optical Motion tracking is a viable solution for a project we are working on.

I wanted to know the best set up, in terms of cameras and software, for optical head motion tracking in an environment that's not really designed for motion tracking. It can be assumed that the individual whose head will be tracked will be stationary (only their head will move). We need 6DOF tracking for the individual's head. The idea is to have enough coordinate data to generate an immersive display output, where the user's head moves around and a virtual camera would rotate or pan accordingly in a 3D scene for example.

Some of the specifics I would be interested in...
*How many cameras would be required for this set up
*Which SDK needs to be used?
*How many markers need to be placed on a user's head for reliable, responsive tracking?

(And also any other advice you can give!)


Thanks,

-Preet

Are you looking to track just the head itself, within the 3D space, or are you looking for more of a facial capture product.

Just looking to track the motion of the head.

Well, you could use a single camera, and the unsupport features of the vector support, within the Optitrack SDK.

You could also go with a 3 camera, rigid body tracking setup.

http://www.naturalpoint.com/optitrack/p ... -body.html

Hi! I have a few more questions.
-What sort of accuracy can I get from this setup (3 camera)?
-Can I use the OptiTrack SDK in conjunction with the Rigid Body SDK?
-Are there any environments in which OptiTrack cameras have interference? Environments that have large amounts of EM noise for example.

1 - Depending on marker size, setup, and calibration, you can get up to sub-millimter

2 - If the cameras are being accessed, and used by the rigid body software, that can't be accessed by another application

3 - EM noise should not be an issue, but anything that is generating IR (heat), could cause interference