I am trying to use HoughTransformCircles method to detect circles of a constant radius. The example applciation HoughCirclesDemo is very helpful in this regard. The application is detecting circles which are not even present on the digital iamges. It arbitrarily selects any point as a circle. Could you please help me what could be the problem in this regard? I will be grateful to you if you could guide me in this regard.
you have to adjust the parameter nMinHits - it is probably to low in your case. Furthermore, for a real application, you would also have to add a hypothesis refinement and verification step after the Hough transform (possibly by least squares regression and computation of a meaningful error measure).
Hello Sir, Thanks for your prompt reply and suggestion.
Shall I use the least squares technique to find the minimum distance between all the circles detected and then remove the ones having minimum distance from each of the circles? Please suggest. Thanks
this forum is for support of the IVT library. I do not teach or provide consulting for machine vision topics here. I hope for your understanding.