Camera Tracking

[jimmyqwerty]

Today we decided to take a close look at programming our camera for the reflective tape, We started by first looking at the 2009 demo tracking code but because we are new to this camera processing programming it was hard for us to understand it, and it was also tracking two colors instead of one. we would like to use java so if anybody had a good reference that would be of use that would be great, or if anybody had very basic code that they would like to share that would be great also.

[Greg McKaskle]

There is a paper about vision at http://decibel.ni.com/content/commun...#/?tagSet=1001 that discusses how the material works and some approaches for processing the images from the camera.

Greg McKaskle

[John Heden]

Great paper! We did something similar last year where we calculated our starting position based on the perceived size of the circle target and their overall ecentricity. It worked extremely well and was much more reliable than having our drive team dial in our starting position.

While it probably does not matter too much this year, I was curious whether any one knew how the 2" retroflective tape would be applied to the 3" target circles ? A couple camera shots from kickoff suggested that the 2" tape would simply be centered in the circle with rounded (to the circle) tops & bottoms (i.e. the whole circle is not covered). This NI paper showed nice 'round' (red) targets. I've found no formal documentation on this subject and would be curious as to the specific retroreflective configuration.

Thanks,

John

[Greg McKaskle]

The construction of the targets wasn't finalized when the paper was written. The circular targets in the paper are slightly smaller and totally covered. Your description matches what I believe the official targets will look like.

A 1x4" strip above and below, and a 2x2.75 strip rounded on top and bottom centered on the circle.

Greg McKaskle