forums.naturalpoint.com

I am trying to optimise the single dot tracking of a reflective ball (roughly the size of a table tennis ball). The problem is I also have infrared LEDs in the vicinity (which can not be removed or obscured). Altering the dot tracking parameters like object_mass does help to favor the reflective ball as the tracked object, but sometimes the ranking switches, making one of the LEDs the tracked object.

Can anyone suggest how I can optimise the dot tracking parameters to ensure the reflective ball remains the tracked object at all times? The individual LEDs are spaced apart (ie not clustered together), and are in close and constant proximity to the reflective ball.

I would be very grateful for any advice.

I have tried tweaking the object_mass type parameters more carefully, and am able to prevent the infrared LEDs in the vicinity of the reflective ball from becoming the tracked object for more of the time. It is not totally consistent though, especially when I increase the distance of the reflective ball and the infrared LEDs from the camera. Could any of the other types of dot tracking parameters help with maintaining the reflective ball as the tracked object? So far my experiments with these other parameters have only made the problem worse.

If tuning the tracking parameters isn't sufficient, perhaps you could perform your own tracking by reading all the objects in the frame and selecting the one with the largest area value.