基于机器视觉的光纤端面几何参数测量研究

研究了一种基于机器视觉的光纤几何参数测量方法。该方法运用数字图像处理算法,首先,读取光纤端面图片,采用Deriche边缘检测算法进行图像分割;然后,利用去伪边缘算法完成对有效边缘的提取及合并;最后,进行曲线拟合,完成对光纤端面几何参数的测量。实验结果表明：该方法操作简便,算法复杂度低,光纤端面边缘拟合精确度较高,数据检测速度快,可以精确高效地检测出光纤端面几何参数的数据,对涂覆层直径的重复测量精度可达0.021μm,涂覆层圆度重复测量精度可达0.043%。

Research on Measurement of Geometric Parameters of Optical Fiber Endface Based on Machine Vision

A measurement method of optical fiber geometric parameters based on machine vision is studied. This method uses digital image processing algorithm. First, read the image of fiber end face and use Deriche edge detection algorithm to segment the image. Then, the effective edges are extracted and merged by using the algorithm of removing false edges. Finally, the geometric parameters of fiber end face are measured by curve fitting. The experimental results show that the method is simple to operate, low algorithm complexity, high fitting accuracy of optical fiber endface edge, fast data detection speed, and can accurately and efficiently detect the geometric parameters of optical fiber end face. The repeated measurement accuracy of coating layer diameter can reach 0.021 μm, and the repeated measurement accuracy of coating layer out of roundness can reach 0.043%.