Automatic detection of chessboard and its applications

Abstract

A chessboard image can be used as a pattern for camera calibration and pose estimation. We propose an algorithm for the automatic detection of a chessboard using one circle as a probe. We first generate candidates using the Lucas-Kanade feature detector algorithm. Finally, we find control points on the chessboard using properties inherent on a chessboard pattern. We suggest that a distinct edge along the border line and black and white sequence on the chessboard can be used as constraints. A circle is used as a probe for checking those properties. We show that the proposed algorithm can be used successfully in camera calibration and pose estimation. Experimental results using images acquired under the variation of scale, pose, and illumination shows the robustness of proposed algorithm.