Precision distortion correction technique based on FOV model for wide-angle cameras in automotive sector

Abstract

This chapter proposes a method for estimating a distortion center to correct the radial distortion that occurs when capturing images with a wide-angle fish-eye lens. In the field of view (FOV) distortion correction model, the error of the distortion center and center of the image increases because it does not estimate the distortion center of a lens separately. This drawback deteriorates the accuracy of distortion correction. Thus, this chapter proposes a distortion correction method using the FOV model and 2D patterns in order to increase the accuracy of the distortion center estimation of a wide-angle lens. To achieve this goal, a distortion curve generated from the FOV model is compared with a straight line, thereby setting the position of the minimum difference between the curve and the straight line as the distortion center. Through this method, the accuracy of the estimation of the center of the distortion that occurs owing the error in the alignment of center points of a lens and the imaging sensor can be improved. This was also verified through experiments. © 2015 Elsevier Inc. All rights reserved.