Does the camera API feed you the image and do you have access to the image.
If I had the image to work with I would have to assume that the pixels are neither totally black or totally white. I couldn't even assume that any of them were absolutely grey unless the webcam API guarantee it.
I'd have to consider a certain sensitivity and require that for a spot to be considered white that it had a pixel value above a certain intensity. Then, assuming I had to scan the entire image to find a spot, I'd scan each line successively until I found a pixel that was at or above my sensitivity setting. I'd then start scanning down from that point until the pixels were under my sensitivity level and find the mid point of the line joining the top and bottom and scan left and right to find where the pixels were under my sensitivity range and let my findings define a rectangle that encompassed the dot and use that information to find a center.
If continual tracking was required from one image to the next I might constrain my scans to the neighborhood of the last spot found unless there was a chance that it would be fast moving.
But if the need to locate the dot didn't occur until something triggered the need.
You might find some information on the process by looking at what Johnny Lee did with Wii tracking. Have a look at this link
http://www.cs.cmu.edu/~johnny/projects/wii/
You might find more information by googling some of the topics you find there. In fact, this link may be even better
http://www.wiimoteproject.com/
Lilith, Night Butterfly
I'm not a programmer but I play one in the office