一种新的基于正交性的摄像机自标定方法  

A New Camera Self-calibration Method Based on Orthogonality

作　　者：潘亚宾[1] 刘国栋[1] 

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《微计算机信息》2011年第4期185-187,共3页Control & Automation

摘　　要：基于摄影几何中的交比不变性、调和共轭理论和正交性与摄像机内参数之间的约束关系,提出一种新的采用正方形模板进行摄像机自标定的新方法。该算法只需在不同方向不同方位拍摄不少于3幅正方形模板的图像,就可以实现摄像机标定。摄像机和模板之间的相对位置可以任意。整个标定过程简单、易于操作,无需知道正方形模板的尺寸大小和空间位置等几何信息。实验表明,该方法精度较高,具有较好的实用性。Based on the invariance of cross ratio, the harmonic conjugate theory and the relation between the orthogonality and the camera intrinsic parameters, a new approach of camera self-calibration using square pattern was proposed in this paper. The method only needed to shoot not less than 3 images of the square planar in different directions. Camera calibration could be achieved. The relative position between the camera and the planar pattern could be arbitrary. The entire calibration process was simple, easy to operate and didn＇t need to know the geometric information such as the size and spatial location of the square pattern. Experimental results show that the algorithm is accurate and convenient. It has better practicalities.

关 键 词：自标定 正交性 正方形模板 摄影几何 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]