A method for the Hough transform detection of circles and ellipses using a 1-dimensional array

A method is proposed for the Hough transform detection of circles and ellipses in binary images using a 1D accumulator array. The method proceeds in two stages: the first stage detects N edge points in an M×M image in both horizontal and vertical scans to construct symmetrical axes using two 1D arrays. The axes cross point determines the object center. The second stage determines the remaining object parameters using the fitting procedure. The present method reduces space and time requirements to O(M log₂ N) bits and O(M²+N²) operations, respectively. Results are given for real and synthetic images with and without noise