Centroid calculation

Post Reply
JeDi
Posts: 41
Joined: Thu Mar 10, 2011 3:59 am

Centroid calculation

Post by JeDi »

Hi,

We are experimenting with different types of active markers, using an infrared L.E.D. A bare L.E.D. works fine, but has a limited angle range towards the camera.

We want to put a diffuser around the L.E.D., but I am not sure about its shape. Obviously a full sphere would be ideal, but the best design for us would be half a sphere. How is the centroid calculated in that case? Will the algorithm detect a circle and properly use the midpoint? Or will it use the midpoint of the light blob?

In the latter case, I'm unsure that the marker detection would work well because the shape of the blob is different from different camera angles.

So, does anyone know the impact on tracking when using half a sphere? It would be about the size of (half) a ping pong ball, and it would be perfect if the recognized 3D position would be in the middle of the base (on the skin).

Greetings,
JeDi
beckdo
Posts: 520
Joined: Tue Jan 02, 2007 2:02 pm

Re: Centroid calculation

Post by beckdo »

The Camera SDK uses 'center of mass' of the blob. I would recommend using a full sphere diffuser.
Post Reply