Automatic checkerboard corner detection combined with gray features and energy minimization (ADGE)

In the camera calibration process, the checkerboard pattern has a wide range of applications. Aiming at the current checkerboard corner detection algorithms that have missed and misdetected situations, this paper proposes an automatic checkerboard corner detection algorithm that combines gray-scale features and energy minimization(ADGE). The ADGE first uses the proposed two sets of gray-level symmetry operators to process the image separately to extract the candidate corner points, and then uses the regional block mode to dynamically search for the local maximum value of the candidate corner points. It further constructs a reaction function to linearly solve the sub-pixel coordinates, and finally extracts all checkerboard corner points and sorts them by minimizing the energy function. Experimental results show that the corner points extracted by the ADGE have no missed and false detection, and the corner reprojection accuracy is 0.1298 pixels. The ADGE can meet the requirements of camera calibration and provide it with high-precision data.