Geometric Correction of Projected Rectangular Pictures

Abstract

This article discusses a method for blindly removing geometric distortions in taken images of rectangular planar patterns (such as maps, pictures, or posters) in the absence of any calibration information or explicit knowledge of the imaging device. Our approach removes lens distortion form single-view images. The method is based on Least-Square-Error (LSE) approximations and Hough transforms. We approximate the lens distortion by considering only lower-order terms of the lens distortion, and additionally a single-coefficient only to model approximately the geometric distortion. The method relies on the assumption that pure radial distortion transforms straight lines into curves, and it exploits plane homography.