Flexible Auto-Calibration and Its Application to Augmented Reality

tion 6. When we have some estimation errors in the autoPractical problems are dealt with for augmented reality calibration, we may have a different perspective view from when the calibration is done through an auto-calibration. what we intended in the embedding step. In Section 7 we Projective motion of the camera is computed and upgradedto investigate the effect of the estimation errors and non-zero Euclidean via an auto-calibration algorithm. Video augmenskewness of estimated cameras. Finally, a concluding remark tation has two steps: embedding graphics objects into the is given in Section 8. real world using epipolar constraint and orthogonality constraint, and rendering the views of the virtual objects. This paper gives details of our approach and the results of real 2 Auto-Calibration from Projective experiments. Reconstruction