A Modified RHT Method for Ellipse Detection Based on Geometric Constraints and Perceptual Grouping

The useless samplings and accumulations of randomized Hough Transform (RHT) are largely reduced with the help of the ellipse´s conic control point constraint and the constraint of the normals at the endpoints of a chord. The fuzzy confidence based on vision perceptual grouping is used to further eliminate false candidate ellipses introduced by the same distorted ellipse. The experimental results show that compared with the other ellipse detection method based on RHT, the method proposed in this paper has the advantage of higher detection speed and accuracy as well as strong resistance to ellipse´s partial occlusion and deformation.