EDCircles: A real-time circle detector with a false detection control
Pattern Recognition
Hi-index | 0.00 |
This paper presents a real-time robust algorithm todetect and accurately locate the circular objects indigital images. The algorithm consists of four steps.First the edge pixels are extracted using Canny edgedetection algorithm followed by a noise removalprocess to remove the non-circle edge points.Afterwards, a direct least square fitting algorithm isdeveloped to calculate radius and circle centerinformation for each edge pixel cluster (a potential arcor a segment of a circle). In third step, a robustcriterion is developed to distinguish the valid arcs frominvalid arcs. Finally, those valid arcs belonging to thesame circle are reassembled and fitting algorithm isrun again to obtain the accurate information of thatcircle. The algorithm is implemented in Visual C++and tested on a laptop powered by an Intel CentrinoDuo CPU at 1.66GHz. The experiment shows thealgorithm’s three advantages. Its speed is fast, about 7images/second for image size of 640X480. It is able toreliably detect full as well as partially-occluded circleobjects even in a noisy environment, specifically 92%correct detection among 174 circles; the achievedaccuracy for radius and center location has reachedsub-pixel level on average.