EVALUATION OF CORRECTION METHODS OF CHROMATIC ABERRATION IN DIGITAL CAMERA IMAGES

This paper reports an experiment conducted to evaluate correction methods of chromatic aberrations in images acquired by a nonmetric digital camera. The chromatic aberration correction methods evaluated in the experiment are classified into two kinds. One is the method to correct image coordinates by using camera calibration results of color-separated images. The other is the method based on the assumption that the magnitude of chromatic aberrations can be expressed by a function of a radial distance from the center of an image frame. The former is classified further into five types according to the difference of orientation parameters common to all colors. The latter is classified further into three types according to the order of the correction function. We adopt a linear function, a quadratic function and a cubic function of the radial distance as a correction function. We utilize a set of 16 convergent images shooting a white sheet with 10 by 10 black filled circles to carry out camera calibration and estimate unknown coefficients in the correction function by means of least squares adjustment. We evaluate the chromatic aberration correction methods by using a normal image shooting a white sheet with 14 by 10 black filled circles. From the experiment results, we conclude that the method based on the assumption that the magnitude of chromatic aberrations can be expressed by a cubic function of the radial distance is the best method of the evaluated methods, and would be able to correct chromatic aberrations satisfactorily enough in many cases