Robust Bias Corrected Least Squares Fitting of Ellipses

Abstract

This paper presents a robust and accurate technique for an estimation of the best-fit ellipse going through the given set of points. The approach is based on a least squares minimization of algebraic distances of the points with a correction of the statistical bias caused during the computation. An accurate ellipse-specific solution is guaranteed even for scattered or noisy data with outliers. Although the final algorithm is iterative, it typically converges in a fraction of time needed for a true orthogonal fitting based on Eucleidan distances of points.