Accurate template-based correction technology for lens distortion

For most off-the-shelf cameras, the lens distortion is inevitable. An accurate template-based calibration method for the lens distortion of an area array digital camera is proposed. A checkerboard template is employed as a standard object in this method and only one image of the template is needed. The mapping relationship in distortion-free imaging is described by a homography matrix and the discrepancy between an observed image point and its ideal location is used to find the distortion center and distortion coefficients. It is straightforward and efficient. For distortion correction, it is known that applying a forward distortion model which transforms the ideal image coordinate (distortion-free) of an image point to its observed image coordinate (with distortion) in a calibration process makes the calibration much more accurate. However, in many cases, it is necessary to calculate the ideal image coordinates for some specified points from their observed image coordinates. Generally, it is often difficult to derive the inverse function of a complex forward distortion model. A distortion-table interpolation method is proposed to accomplish the transformation. Both computer simulation and real data are used to validate the proposed methods and good results are obtained. (C) 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). [DOI: 10.1117/1.OE.51.10.103602]