中文摘要：

多摄像机是解决结构光测量系统中视场小和自遮挡等问题的有效方案,但同时由于摄像机分布广、无公共视场,存在全局标定困难的问题。提出了一种基于条纹投影的多摄像机标定方法。加入辅助设备投影仪并向视场内投影大幅面的条纹图案,借助相位算法建立起全局相位场,反求景物的投影仪虚拟图像。利用平面靶标标定摄像机的内参和投影仪的外参获取靶标平面的移动信息,进而求解相邻摄像机的位置参数。最后的优化过程中,再根据相位值对相邻摄像机的图像完成亚像素匹配,通过最小化重建偏差对系统参数进行迭代优化。在搭建的多摄像机三维测量系统上进行试验,结果表明本文的多摄像机标定和测量方案切实可行。

英文摘要：

Multi-camera is an effective solution to address the small field of view and self-occlusion problem in the structured-light system. Due to the wide distribution of cameras and without the public field, there is a global calibration difficulty. A multi-camera calibration method is presented based on fringe projection. A projector, auxiliary equipment, is introduced, and the large-format fringe image is projected within the field of view. A global phase-field with phase algorithm is established, and the virtual image of the scene is inversely solved in the projection plane. In order to obtain the moving information of target plane, a two-dimensional target is used to calibrate the internal parameters of camera and external parameters of the projector. Therefore, the positional parameters of the adjacent camera can be solved. In the final optimization process, sub-pixels match the image of the adjacent camera based on the phase values, and the system parameters are iteratively optimized by minimizing the reconstruction deviation. The experiment on multi-camera structured-fight system shows that the multi-camera calibration and measurement is feasible.