Camera Calibration Method For Stereo Measurements

Abstract

Camera calibration is an important and sensitive step in 3D environment reconstruction by stereovision. Small errors in the initial estimation of the camera parameters (especially in the estimation of the principal point) could rise to high errors in the 3D measurements, which are increasing with the working distance. There are many general-purpose calibration methods for calibration of individual cameras. We propose a method for redefining the results of such methods by inferring stereo information from a controlled 3D scene. The parameters of the two cameras composing a stereovision system are recalibrated together by minimizing the depth reconstruction error of the control points from the reference scene.