Multi-circle Detection Algorithm Based on Symmetry Property
Lianyuan Jiang, Peihe Tang, Yingjun Zhu, Jianbing Jiang, Yalan Zhang · International Journal of Applied Mathematics & Statistics/International journal of applied mathematics and statistics · 2013
The randomized Hough transform (RHT) and the randomized circle detection algorithm (RCD) are commonly used for circle detection, but the use of these two approaches for images with high percent of noise will cause a large amount of time in calculation. To detect circles fast in multi-noise images, the present paper proposes a multi-circle detection algorithm on the basis of the symmetry property of circle. We first find 4 symmetric points in edge image, calculate a circle parameter from the coordinate values of three of the symmetric points, and then affirm whether the correspondent circle is a true circle or not. In this method, we need not to know the number of the circles in the image in advance, and the detecting time is not affected by the number of circles in the image or the size of the circle. Experimental results demonstrate that our algorithm has the advantage of high detection accuracy and of faster speed in circle detection compared with RHT and RCD.