[logicalhippo]

This is a SERIOUS problem that at the moment I'm quite concerned about. (the two tetra problem) - basically what it comes down to is you're going to have to do something tricky. The camera detects the center of mass of the calibrated color in the image, which means that if two tetras of equal distance are in the field of vision, the camera will return the point between them.

One thought I had was you could start your robot facing paralell to the wall, set the x-range of the view window to a really small value (10-20), and then slowly rotate the robot until it sees a tetra. because of the small window, we can be reasonably sure it is only one tetra. I haven't tested this yet.