There are three flavors of Hough transform. The first (original) was used to detect straight lines in bubble chamber images. The technique parameterizes pattern space, then projects all points of interest (typically edges found via Canny edge detection) into the parameter space. Each point votes for every possible shape parameterization it could belong to. The resulting accumulator array is then thresholded and non-maxima suppressed, keeping only local minima. Indices of the bins in the accumulator correspond to shape parameterizations. By looking at bins still populated after thresholding and non-maxima suppression, it is possible to extract the parameters of all shapes of a given type in an image.
Hough Transform
This entry contributed by Brendan Drew
Explore with Wolfram|Alpha
References
Shapiro, L. G. and Stockman, G. C. Computer Vision. Englewood Cliffs, NJ: Prentice-Hall, 2001.Trucco, E. and Verri, A. Introductory Techniques for 3-D Computer Vision. Englewood Cliffs, NJ: Prentice-Hall, 1998.Referenced on Wolfram|Alpha
Hough TransformCite this as:
Drew, Brendan. "Hough Transform." From MathWorld--A Wolfram Web Resource, created by Eric W. Weisstein. https://mathworld.wolfram.com/HoughTransform.html