I am looking to use HoughCircles to find circles where I have a strong prior on where the circle center(s) are. Is there any easy way to incorporate that information into the Hough Circles algorithm? i.e. I mostly want to sweep radius values and find the best radius / radii.
I’m not a big fan of the Hough transform. I’m probably not well informed enough to state an opinion, but it seems to me to be the bubble sort of feature detection. It’s popular because it’s easy, not because it’s good.
With that out of the way, I’m going to suggest you don’t try to incorporate your known circle center into Hough circles, although that might actually be quite easy and effective.
I have used this in the past:
But that requires more work to implement. Not a ton, but maybe not the answer you are looking for. It depends on what level of accuracy you need, and if it’s high I would question how confident you are in your center point. For my application I was getting very repeatable subpixel results. I don’t remember the details, but probably on the order of < 0.05 pixel variation from repeated imaging of the same circle.
If you want to use Hough circles with a fixed center point, I’d take a look at the source code and try to figure out how to reduce the search space to just the radius. It would be a lot faster, and/or you could get higher precision radius estimates by using fiber discretization steps for the radius.