Detection and model analysis of circular feature for robot vision

A novel detection and model analysis of circular feature with error compensation is addressed in this paper for robot vision. We use an efficient method of fitting ellipses to data points by minimizing the algebraic distance subjected to a constraint that a conic should be an ellipse, and solve the ellipse parameters through Lagrange multiplier method. Using closed form solution, the 3D position and orientation of a circular feature with known radius can be obtained. Since the center of projected circle does not correspond to the ellipse center, we compensate for the projection error between them. Experimental results show that our method is robust, efficient and accurate.