视觉导航中的单目摄像机标定改进算法

在基于单目视觉的智能车自主导航系统中，摄像机标定是智能车实现安全准确视觉导航的前提和关键。研究摄像机理想模型和实际模型，提出一种改进的标定方法。结合线性模型和非线性模型的优点，在求解摄像机内外参数的过程中，先采用线性模型标定摄像机的一部分参数，进一步考虑简化畸变模型，将非线性方程组转化为线性方程组迭代求解，最终获得摄像机全部参数。该方法既保证了标定精度，又简化了复杂的摄像机实际模型。实验结果表明该方法能满足视觉导航要求。

Improved algorithm of monocular camera calibration for vision navigation

One camera is used in the autonomous navigation system of intelligent vehicle, and camera calibration is the key and precondition of the intelligent vehicles correct and safe navigation. Through the study of ideal model and actual model of camera, an improved calibration method was proposed. In the process of obtaining internal and external parameters of the camera, this algorithm combined the advantages of both linear and nonlinear calibration, which calibrated parts of the parameters of camera firstly and then considered simplifying the distortion model. The nonlinear equation systems can be changed into linear equation systems. Additionally, all the parameters of camera were acquired through many times iteration. On the one hand, it can ensure the calibration accuracy. On the other hand, it reduces the complexity of the cameras actual model. The experimental results show that the method can meet the requirements of vision navigation.