A New Sub-Pixel Detector for X-Corners in Camera Calibration Targets

Abstract

X-corner patterns are most widely used in camera calibration. In this paper, we propose a new sub-pixel detector for X-corners, which is much simpler than the traditional sub-pixel detection algorithm. In this new algorithm, the pixel position of X-corner is firstly detected by a new operator. Then a second order Taylor polynomial describing the local intensity profile around the corner is educed. The sub-pixel position of X-corner can be determined directly by calculating the saddle point of this intensity profile. Neither preliminary intensity interpolation nor quadratic fitting of the intensity surface is necessary, which greatly reduces the computation load of the detection process. Furthermore, computer simulation results indicate that our new algorithm is slightly more accurate than the traditional algorithm.