A stratified self-calibration method for a stereo rig in planar motion with varying intrinsic parameters

Self-calibration for imaging sensors is essential to many computer vision applications. In this paper, a new stratified self-calibration method is proposed for a stereo rig undergoing planar motion with varying intrinsic parameters. We show that the plane at infinity in a projective frame can be identified by (i) a constraint developed from the properties of planar motion for a stereo rig and (ii) a zero-skew assumption of the camera. Once the plane at infinity is identified, the calibration matrices of the cameras and the upgrade to a metric reconstruction can be readily obtained. The proposed method is more flexible than most existing self-calibration methods in that it allows all intrinsic parameters to vary. Experimental results for both synthetic data and real images are provided to show the performance of the proposed method.