A new approach for distorted image restoration by distance-based correction and cubic spline interpolation

Abstract

Fish-eye lenses are used more and more widely in the field of photography and computer vision nowadays. However, images captured with these lenses suffer from high barrel-type spatial distortion, so severely errors are brought into the camera calibration, feature extraction, parameter calculation and following analyses. This paper proposes a new and effective method to correct the high distortion. With the improved model in this paper, the degree of correction changes in accordance with the distance from the center pixel location. And the inverse mapping from corrected images space to the distorted images space based on the cubic spline interpolation method is taken to eliminate the vacant pixels in the corrected images caused by forwarding mapping. The final correction results are satisfactory by using this improved method.