Automatic focal length estimation as an eigenvalue problem
A robust linear method for auto-calibration of a moving camera from image sequences is presented. Known techniques for auto-calibration have problems with critical motion sequences or biased estimates. The proposed approach uses known linear equations that are weighted by variable factors. Experiments show, that this modification reduces problems with critical motion sequences and that the estimates are not biased. Therefore, the proposed approach is more robust and achieves a higher estimation accuracy.