I would recommend just using iRvision. It is plug and play with the robot and has the built in software you calibrate it to the robot.
The manual is large but pretty good in my opinion. I suggest learning on your own then consult the manual with keyword searches or looking at the table of contents as needed. One you get more familiar with programming Fanuc, then you can start reading more of the manual and understand it better.
Thanks for the reply HawkME! Part of the work to be done with the robots is to try some in-house developed visual object classifiers in the lines, this is the main reason behind this question. I have very limited knowledge of RVision and its capabilities, maybe we could couple our recognition system directly with it instead of using something entirely different.