A novel approach for automation of stereo camera calibration process

The problem of stereo camera calibration has been studied for over many years by numerous researchers. A crucial task in this process is to discover the transformation between 3D world coordinates and 2D pixel image coordinates of image. The growth of the number of different applications of stereovision systems has led to specialization of developed calibrations algorithms. Nowadays, various calibration objects and selfcalibration techniques are used. This paper presents a unique automatic calibration system for a stereovision system for inspection of specimen surface under fatigue tests. In order to allow analysis of surface in both a micro and macro scale, the system has been equipped with cameras with motorized focus and zoom lenses. The proposed calibration system is based on mechatronic framework which allows the use of a set of 2D plane calibration targets with varying size of region of interest. Such a solution allows automation of calibration process and guarantees repeatability of results with an assumed error.