Conjugate Epipole-based Self-Calibration of Camera under Circular Motion

摘要

In this paper, we propose a new method to self-calibrate a camera with constant internal parameters under circular motion. The basis of our approach is to make use of the conjugate epipoles which are related to camera positions with rotation angles satisfying the conjugate constraint. A novel circular projective reconstruction is developed for computing the conjugate epipoles robustly. It is shown that for a camera with zero skew, two turntable sequences with different camera orientations are needed, and for a general camera three sequences with different camera orientations are required. The performance of the algorithm is tested with real images.