提出了一种基于图像的保偏光纤尾纤定位技术, 研究了图像采集模块, 设计了粗检测与精检测相结合的边缘检测策略。首先采用自适应循环变化阈值算法对尾纤边缘进行初步检测, 增强了光照鲁棒性, 其中提出起始点自寻的区域搜索法能够准确提取目标边缘点; 然后通过基于三次多项式函数拟合的亚像素检测方法提高了边缘检测的精度。搭建了保偏光纤尾纤自动定位系统, 并进行了图像处理实验和旋转定位实验。实验结果表明, 本文的尾纤定位技术能够实现0.06°的保偏光纤尾纤旋转定位精度, 具有重要的工程应用意义。

A positioning method based on image processing for PMF pigtail was proposed in this paper, and an image acquisition module and an edge detection strategy combining rough detection and precise detection were researched and designed. For rough detection, an adaptive cyclic change threshold segmentation is applied to enhance the robustness to light intensity, and a revised area search algorithm based on initial point self-finding is found to properly extract the target edge points in preliminary detection. For precise detection, the method further improves the detection accuracy by a subpixel edge detection based on cubic polynomial function curve fitting. In order to verify this method, an automatic positioning system is established. Several results of image processing and positioning experiments show that this paper can achieve the rotational accuracy less than 0.06°.