【摘要】 提出了一种新的基于主动视觉系统的线性摄像机自定标方法 .所谓基于主动视觉系统 ,是指摄像机固定在摄像机平台上以及摄像机平台的运动可以精确控制 .该方法的主要特点是可以线性求解摄像机的所有 5个内参数 .据作者所知 ,文献中现有的方法仅能线性求解摄像机的 4个内参数 .当摄像机为完全的射影模型时 ,即当有畸变因子 (skew factor)存在时 ,文献中的线性方法均不再适用 .该方法的基本思想是控制摄像机做 5组平面正交运动 ,利用图像中的极点 (epipoles)信息来线性标定摄像机 .同时 ,针对摄像机做平移运动时基本矩阵的特殊形式 ,该文提出了求基本矩阵 (fundamental m atrix)的 2点算法 .与 8点算法相比较 ,2点算法大大提高了所求极点的精度和鲁棒性 .另外 ,该文对临近奇异状态 (即 5组平面正交运动中 ,有两组或者多组运动平面平行 )作了较为详尽的分析 ,并提出了解决临近奇异状态的策略 ,从而增强了该文算法的实用性 .模拟图像和真实图像实验表明该文的自标定方法具有较高的鲁棒性和准确性更多还原

【Abstract】 In this paper, a new active vision based linear camera self-calibration technique is proposed. By an active vision system, the camera is rigidly mounted on a camera platform and the platform can be controlled to undergo precise motions. The novelty of this new calibration technique is that it can determine linearly all the five intrinsic parameters of the camera. To our knowledge, techniques reported in the literature up to now can cope with linearly only four of the five parameters. In other words, when the camera is of a complete projective model, i.e., when the skew factor is non-zero, such techniques become invalid. The basic principle of the new calibration technique is to control the camera to undergo at least five sets of orthogonal planar motions, then the camera’s intrinsic parameters can be linearly determined via epipoles. In addition, since the fundamental matrix must be of an anti-symmetric one if the camera’s motion between the two images is a pure translation, a 2-point algorithm, rather th an the traditional 8-point algorithm, is used to determine the fundamental matrix. The 2-point algorithm is proved to be a great contributor t o the substantial increases of the robustness and accuracy of the final calibration results with a large number of experiments on synthetic and real images.更多还原