Research on vision-based waveguide alignment of optical fiber

Tao Geng,YuZhuo Wang,HaiBin Wu

DOI: https://doi.org/10.1109/mic.2012.6273332

2012-05-01

Abstract:Optical fiber waveguides alignment is the key technique of optical fiber tapering system. In this paper, we propose a waveguide alignment method based on machine vision measurement technology to accomplish precise alignment of different types of microstructure fibers. An array charge-coupled Device (CCD) is employed as image sensor and such as by analyzing the performance of different kinds of light source and CCD lens, a high accuracy microscopic measurement system has been constructed. On the basis of parameter optimization, the auto-focus system has been performed in terms of the point sharpness method. Canny operator with good noise immunity is chosen as a full pixel edge detection operator to pick up the fiber edge. After calibration, the deviation between the two aligned fibers can be calculated and obtained, and then the precise alignment of fibers can be achieved by adjusting micro motion actuator. The alignment error between fiber-cores is less than $2{\mbi \mu}{\bf m}$. The fiber waveguide alignment method has advantages of good stability, fine real-time and high alignment accuracy, which can provide a foundation to improve the performance of fiber tapering system.