A new camera calibration based on vanishing point

Camera calibration is an important issue in machine vision. This paper proposes an improved calibration based on vanishing point to determine camera intrinsic parameters as well as its pose. According to projective geometry constraint, we compute all camera parameters directly by two vanishing points. Compared with existing methods based on vanishing point, the proposed approach decreases variables without increasing the number of images. As a result, the proposed approach reduces errors and simplifies calibration process. In addition, the proposed approach reduces constrains and avoids the correspondences between images and space planes so as to be flexible. Simulations and real image experiments validate the proposed approach and show that it is accurate and robust to noise.