最近，张正友提出用平面模板两步法来标定摄像机。该方法能高精度地标定摄像机，且简便易行，可以说是两步法标定的代表，但其算法模型仅考虑了镜头径向畸变，其对切向畸变较大的场合，如鱼眼镜头成像，则不适用。为此，通过改进其摄像机模型，提出了一种改进的两步法，该方法先用图像中心附近点求取初值，由于图像中心附近点畸变很小，故求取的初值能很好地逼近准确值，然后采用一种基于内部映射牛顿法的子空间置信域法通过求精来得到所有参数。仿真实验和真实实验的结果表明，该方法的标定精度要优于张正友的方法，鲁棒性也更好。

Zhengyou Zhang has proposed a Two-Stages camera calibration method by viewing a plane recently. It can calibrate camera easilywith high accuracy and is considered as a presentation of Two-Stages method. But only radial distortion is considered in his camera model, and when the tangential distortion is distinct in some lenses, such as in fish-eye lenses, the camera model will be unsuitable. Therefor his camera model is improved in this paper and an improvedTwo-stages method is proposed. Through the proposed method the initial parameters are solved by using the points near the image plane center. As the distortion is very litter nearthe image plane center, the solved initial parameters can be very close to exact ones, then, a subspace trust regionmethod based on the interior-reflective Newton method is applied to solve all the camera parameters accurately. Both computer simulation and real data have been used to test the proposed technique and very good results have been obtained. Compared with Zhang’s method, the proposed method is more accurate and it works better robustness.