The Camera Calibration Approach of Pipeline Detecting Robot

Abstract

Camera calibration is a key technology in computer vision field. Many kinds of distortions are existed in ordinary CCD camera. Among these distortions the radial distortion is more serious. A camera self-calibration algorithm based on active vision taking account of one-degree radial distortion is proposed in this paper in order to make the calibration more quickly effective and accurate. Based on pinhole camera model, considered the camera lens distortion effects of radial and tangential distortion effect, it reduces the iterative times and avoids errors that come from decomposing parameters in linear calibration. The results show that the algorithm for high precision and with a certain degree of robustness, achieve fast and accurate calibration of vision cameras for weld inspection robot.