Fast detection of multi-circle with randomized Hough transform

Abstract

Randomized Hough transform (RHT) is an effective method for circle detection. But when dealing with multi-circle complex images, the random sampling will bring lots of invalid accumulations and result in a large number of calculations. In this paper, by selecting three points of the candidate circle, a fast detection algorithm of multi-circle with randomized Hough transform is presented. Experimental results demonstrate that the proposed scheme can quickly detect multiple circles, and has a strong robustness.

Keywords

Image Point / Synthetic Image / Neighborhood Information / Strong Robustness / Circle Parameter

References

[1] XuL., OjaE., KultanenP.. Pattern Recognition Letters, 1990, 11: 331

[2] XuL., OjaE.. CVGIP: Image Understanding, 1993, 57: 131

[3] WalshD., RafteryA. E.. Pattern Recognition, 2002, 35: 1421

[4] JiQ., XieY.. Pattern Analysis and Applications, 2003, 6: 55

[5] LiZ.-q., TengH.-f.. Journal of Computer-Aided Design and Computer Graphics, 2006, 18: 27

[6] XuL.. Pattern Recognition, 2007, 40: 2129

[7] WangK.-j., PingZ.-l., HaiY.. Journal of Optoelectronics·Laser, 2008, 19: 420

[8] JiangL.-y., SuQ., ZhuY.-j.. Computer Engineering and Applications, 2009, 65: 163

[9] ChenT.-c., ChungK.-l.. Computer Vision and Image Understanding, 2001, 83: 172