A Circle Detection Algorithm Based on Mathematical Morphology and Chain Code

Circle is often used in application of computer vision. It is the most essential and most important task to detect circle in measurement based image. Hough transform is used widely in circle detection. But when lens and circle surface are non-parallel because of objective reason, leading to obtain the actual image is not a standard round, the conduct of this detection algorithm will be extremely difficult, or even not possible. So it can not be used in real time detection. This paper proposes an algorithm of circle detection which is based on mathematical morphology and chain code. It uses digital image processing techniques for binarization image and morphology method for completing edge contour extraction. Then it determines the circularity index by chain code of the edge of the image. Finally, it compares the circularity index with the preset threshold, so as to calculate the centroid. Experimental results show that this algorithm is simple and effective, and is sub-pixel accuracy.