Self-calibration algorithm under varying cameras using the linear projective reconstruction method

We present a practical self-calibration algorithm that only requires a linear projective reconstruction. Recently, many self-calibration algorithms that use only the information in the image have been proposed. But most algorithms require bundle adjustments in the projective reconstruction or in the nonlinear minimization. We overcome the sensitivity of the self-calibration algorithms due to the image noises by adding another constraint on the position of the principal point. We also propose a linear initialization method based on the property of the absolute quadric. Experimental results using real and synthetic images demonstrate the feasibility of the proposed algorithm.