Abstract A commercial stereo light microscope, such as the OLYMPUS SZX7, has nonnegligible lens distortion for three-dimensional microscopic image correlation. To tackle the problem, an effective calibration method of using a stereo-vision system is proposed. It is implemented using non-coplanar feature points with weighted radial constraints, where the weights are incorporated into the objective function to estimate the internal and external parameters of the vision system iteratively. In the calculation of the distortion coefficients, a non-parametric distortion model expressed by a bicubic spline surface is used. The iterative convergence of the objective function and the influence of the attitude number of the calibration target on the calibration results are analysed, and verified with a digital stereo light microscope established with OLYMPUS SZX7. Both shape and displacement measurements have been performed to demonstrate the proposed method. The experimental results show that the proposed method can achieve higher accuracy compared with Zhang's classical method and the unweighted two-step method.

中文翻译：

---

基于具有迭代加权径向对准约束的非共面特征点的立体显微镜校准

摘要 商用立体光学显微镜，例如 OLYMPUS SZX7，对于三维显微图像相关性具有不可忽略的镜头畸变。为了解决这个问题，提出了一种使用立体视觉系统的有效校准方法。它使用具有加权径向约束的非共面特征点来实现，其中权重被合并到目标函数中以迭代地估计视觉系统的内部和外部参数。在畸变系数的计算中，使用了由双三次样条曲面表示的非参数畸变模型。分析了目标函数的迭代收敛性和标定目标姿态数对标定结果的影响，并使用 OLYMPUS SZX7 建立的数字立体光学显微镜进行验证。已经进行了形状和位移测量以证明所提出的方法。实验结果表明，与张的经典方法和未加权的两步法相比，所提出的方法可以获得更高的精度。