We propose an easy-to-implement and accurate calibration method for large-scale stereo-digital image correlation (stereo-DIC). First, the intrinsic parameters of each camera are separately calibrated using a regular-sized phase target. Using phase-shifted circular fringe patterns that feature the advantage of being insensitive to the defocus, the phase target can be placed in the close range of each camera. Then, scale-free extrinsic parameters are computed from the epipolar geometry, which can be easily retrieved from DIC registration of homologous point pairs in the stereo images of a test specimen surface. These intrinsic and scale-free extrinsic parameters are used as the initial guess for further optimizing the calibration results. Further, by measuring objects with known physical size, the scale information of the stereo-DIC system can be determined. The metrological performance of the proposed method is evaluated by large-scale shape reconstruction and deformation measurement tests. The obtained results demonstrate that the proposed method is a viable solution for the large-scale 3D deformation measurement tasks.

我们提出了一种易于实施且准确的大规模立体数字图像相关（stereo-DIC）校准方法。首先，使用常规大小的相位目标分别校准每个相机的内在参数。使用具有对散焦不敏感的优点的相移圆形条纹图案，可以将相位目标放置在每个相机的近距离范围内。然后，从对极几何计算无标度的外在参数，可以很容易地从试样表面立体图像中同源点对的 DIC 配准中检索这些参数。这些内在和无标度的外在参数用作进一步优化校准结果的初始猜测。此外，通过测量已知物理尺寸的物体，可以确定立体 DIC 系统的尺度信息。通过大规模形状重建和变形测量测试来评估所提出方法的计量性能。获得的结果表明，所提出的方法是大规模 3D 变形测量任务的可行解决方案。