To implement accurate metric measurement within a large field of view with common off the shelf cameras

the image distortion caused by wide angle lenses should be corrected. A simple and precise method for calibrating the radial distortion is proposed. Any planar pattern in which the feature points form a grid can be used to obtain the calibration image. The proposed method first constructs an ideal image based on the positions of feature points extracted from the distorted image. Then the radial coordinates of the distorted and ideal feature points are calculated under the hypothesis of given center of distortion

and the relation function between the ideal and distorted radial coordinates is obtained by solving linear equations or by polynomial fitting. The set of parameters with the minimum mean square error is regarded as the calibration result. One merit of the proposed method is that both the undistorted to distorted and distorted to undistorted expression can be obtained at the same time by the same way. Moreover

the inverse power proportion relation between calibration result and resolution of rectified image is derived

and the influence of the order of distortion function on rectified image is discussed. Experiment results show that 4th or 5th order function would result better outcome.