关键词: 椭圆检测, Hough变换, 椭圆中心估计, 聚类

Abstract: A fast automatic detection of ellipses is very important for the understanding of some images.In order to overcome the excessive time and storage requirements associated with the Hough transform，we introduce an improved algorithm.This algorithm uses geometric symmetry to estimate the potential centers of ellipses.It also takes the advantages of major axis of an ellipse to find center and orientation parameters of the ellipse，and uses an one-dimensional accumulator array to accumulate the length of minor axis of the ellipse.By using data clustering，it classifies all detected virtual ellipses into their corresponding real ellipses.Experiments with both synthetic and real images indicate the correctness and effectiveness of the proposed algorithm.

Key words: ellipse detection, Hough transform, ellipse center estimation, clustering