Ellipse Detection Method based on the Advanced Three Point Algorithm

In this paper, we propose a fast ellipse detection method using the geometric properties of three points, which are the components of an ellipse. As many conventional ellipse detection methods carry out the detection using five points, a random selection of such points requires much redundant processing. Accordingly, in order to search for an ellipse with minimum number of points, this paper uses the normal and differential equation of an ellipse which requires three points based on their locations and edge angles. First, in order to reduce the number of candidate edges, the edges are divided into 8 groups depending on the edge angle, and then a new geometric constraint called quadrant condition is introduced for the reduction of noisy candidate edges. Clustering is employed to find prominent candidates in the space of some ellipse parameters. Experiments through real images show that our method satisfies both the reliability and detection speed of ellipse detection.