Ellipse fitting algorithm based on subspace projection technique
Yang Zhonggen · Journal of Shanghai, Maritime University · 2006
By means of twice subspace projections,fitting error vector is orthogonally decomposed as three components.The optimal estimation of the ellipse parameter vector is given by simultaneously minimizing the norms of three components.It results in a total optimization process in which the norm of one component,i.e.orthogonal projection of fitting error vector in centralized order-one subspace,is minimized subject to the constraints that the other two components, i.e.the error average and the projection of fitting error vector in centralized order-one subspace,are equal to zeros.The optimal process can be combined with singular value decomposition and progressively completed.The theoretical analysis and experimental demonstration prove that the new algorithm is fast and exact,its anti-noise ability is strong and its rate of successfully fitting is high.