SELF-CALIBRATION FROM 4 VIEWS WITH POSSIBLY VARYING INTRINSIC PARAMETERS

In this paper we address the problem of evaluating how accurate the Eu clidean reconstruction using a self-calibration method can be. We focu s on the self-calibration of a binocular head-eye system from four vie ws with varying intrinsic and extrinsic parameters. The only informati on provided to the algorithm is the fundamental matrices [1] and the p oint correspondences between the four views. The main improvements of this article are: (a) to take account of intrinsic and extrinsic param eters change; (b) to exploit the information of the cross-corresponden ces to improve the Euclidean reconstruction.