An Automatic Algorithm for Accurately Locating Calibration Markers on a Chessboard

Locating the calibration markers of a chessboard is a prerequisite step to camera calibration. The efficiency and accuracy of this step has a crucial influence to the final results. This paper presents an automatic method for detection and location the internal corners of black and white squares by using Multi-Radon Transform (MRT) and image segmentation methods. The whole procedure mainly includes two steps: (1) rough positioning of the markers by using a Triple Radon Transform algorithm (TRT); and (2) precision positioning of the markers by using the Radon Transform and image segmentation comprehensively. Using this method, experiments show that a fully automatic camera calibration could be achieved, and the proposed approach is robust in dealing with different pictures that vary in different circumstances.