Self-calibration is the computation of metric properties of the cameras and the scene from a set of uncalibrated images. This is different from conventional calibration where the camera calibration matrixKis determined from the image of a known calibration grid or properties of the scene

such as vanishing point of orthogonal directions. Instead

in self-calibration the metric properties are determined directly from constraints on the internal or external parameters. Camera calibration is essential to many computer vision applications. In practice this often requires complicated calibration procedures to be carried out regularly. In the last few years a lot of work has been done on self-calibration. It has been shown that a metric calibration was possible based on the rigidity of the scene. Based on the pin-hole camera model

a new self-calibration method is presented in this paper. We can use the char- acteristic of 3-point perspective projective

vanish points and the orthogonal vector to get a set of equations and then all the intrinsic parametersαx

αy

u0

v0can be solved with high accuracy. The feasibility of the approach is illustrated on synthetic images.