Testing the Verticality of Slow-Axis of the Dual Optical Fiber Based on Image Processing

The polarization dual fiber collimator is one of the important passive devices, which could be used to transform divergent light from the fiber into parallel beam or focus parallel beam into the fiber so as to improve the coupling efficiency of fiber device. However, the quality of double-pigtail fiber used for assembling collimator directly impacts on the performance of fiber collimator. Therefore, it is necessary to detect the quality of double-pigtail fiber before collimators has been packaged to ensure the quality of fiber collimator. The paper has pointed out that the verticality of two slow-axis of the double-pigtail fiber is the major factor to affect the quality of fiber collimator. With "Panda"-type double-pigtail fiber as the research object, a novel method to detect the verticality of slow-axis is proposed. First, with the red light of LED as background light source, the clear images of the cross-section end surface of double-pigtail fiber has been obtained by micro-imaging systems; Then after a series of image pre-processing operations, the coordinate information of "cat\´s eye" light heart can been extracted with the image centroid algorithm; Finally, the angle between the two slow axes of the double-pigtail fiber can be calculated quickly and accurately according to pre-established mathematical model. The detection system mainly consists of CCD, microscopic imaging device, image board and image processing software developed with VC++. The experimental results prove that the system is both practical and effective and can satisfy the collimator assembly precision demands.