一种有效的基于消失点的摄像机自标定方法

摘要

An effective camera self-calibration method based on the vanishing points is proposed in this paper. In order to calculate the ef-fective vanishing points with less images,firstly,several images are taken in different locations with proper angles,two classes of orthogo-nal line segments,where in the line segments are parallel with each other in the 3D space,are detected based on the methods of J-linkage and LMedS from the images. Secondly,the vanishing points are obtained with MLE,and four pairs of orthogonal vanishing points are se-lected to calibrate the camera's intrinsic parameters. Finally,in order to reduce the influence of the errors of the vanishing points,the method of condition adjustment with parameters is introduced to improve the accuracy of the camera's intrinsic parameters. The experi-ments are conducted under real conditions,the results show good precision and stability of the calculated intrinsic parameters using six im-ages,that proves the effectiveness of the proposed calibration method.%文中提出了一种有效的基于消失点的摄像机自标定方法。为了使用较少幅图像获得有效的消失点估计，文中首先选择合适的角度在不同位置拍摄若干幅图像，然后使用J-linkage和LMedS的算法检测每幅图像中在场景内两个正交的平行线段组，再使用MLE最优化方法求取消失点，任意选择四对正交消失点标定摄像机内参。为了减少消失点估计误差对标定摄像机参数精度的影响，文中进一步引入附有参数的条件平差方法，调整参数的估计值，提高摄像机内参的精度。实验采用真实环境中的图片，实验结果表明只要6幅图像就可以得到摄像机准确的内参，具有较好的精度和稳定性，说明文中标定方法是有效的。