A Modified Algorithm for Incomplete Corner Identification of Chess Board Target

For the problem of local edge corner failure detection by Hessian operator in the process of the chess board corner detection, a new modified algorithm is proposed to recover the corner recognition. In this algorithm, firstly, based on coordinate extremes method, the chess board poles and the envelope corners on the edge was identified automatically, and the three feature points of head, tail and middle corners on two edge lines were extracted, on which all envelope corners can be detected. After that, the vanishing points of the edge lines were obtained by the use of cross ratio invariance, thus the no-detected pole can be calculated according to the principle of vanishing point. Finally, all the corner coordinates were obtained by the reprojection from the homography relationship between the image coordinates and the world coordinates of the four poles. The experimental results show that this method has strong robustness and can be realized automatically with compared to the common detection methods, so it is of great practical value for camera calibration.