I have looked at a few methods for calibration of a simple omnidirectional camera using a catadioptric lens. I have found two that look promising. These are here:
First, I want to explain that I have little understanding of computer vision and am rusty when it comes to C (# and ++ too). But, I am sure that this will work for our purposes when we get there.
I will try to test at least one of these out in the next month or so. This will make this project portion development to be only a year long. I will try to post some images of getting everything set up.
Things I want to say but are off topic:
Originally, I was going to link to a calibration that used Matlab, but decided it probably shouldn't be used due to it being not remotely open source and being dependent on Windows.
Also, I want to say any search for catadioptric systems will most likely result in an abundance of telescope and telescope lens information. If you want to do your own fishing for information on such designs, be sure to include the word omnidirectional in there...unless you are looking for telescopes.
I had mentioned somewhere on this page that I waned to put this project on the back burner for some time while I purchase and use a dual camera rig--I have realized I don't have money for the cameras, so I am moving back to using this vision system.
Once this is working we can start on the localization and mapping. I hope that won't take a year to get to.
9/29/18