Hi,
The basic concept is about storing the (x, y, r)coordinates in an array (accumulator array - temporary array). The details on the accumulator you can find here: http://me.umn.edu/courses/me5286/visi.... Since this is ineffective, OpenCV uses the Hough Gradient Method.
I think it is also helpful to note that the term "accumulator" refers to its function of accumulating votes (often weighted votes) for a specific geometric configuration. In the case of a circle, all edge responses are voting for those particular circles with centers and radii which that edge response might be a part of. As simon2k notes above, the basic Generalized Hough concept is not very effective for circles and the algorithm has been modified to be more efficient. But the general concept is to accumulate votes over the possible variations in the parameter space.
Asked: 2018-08-27 00:35:50 -0600
Seen: 426 times
Last updated: Aug 27 '18
