Optical Fiber Defects Detection Based on Gabor Filter

In this study, an optical fiber defect detection system based on Gabor filter is proposed. Firstly, in order to eliminate the effect of background texture and enhance flaw characteristics, a set of filters including Median filter and Gabor filter are used, defect features can be preliminary extracted after morphological operations and threshold segmentation. Secondly, the filtered image is divided into small and identical size sub-blocks. Depending on the number of defects features each sub-block contained, divide those sub-blocks into defective blocks and non- defective blocks. Finally, due to a defect often shown as an area, it can be seen as a set of defective blocks. Defective areas are selected from a domain, which consist of a defective block and its neighborhoods. After a sifting based on the grayscale of the original image and the shape and location feature of the defective areas, the image has been judged. The experimental results indicate that the proposed method can be applied in the non- destructive and effective inspection of optical fibers and provides a feasible solution for optical fiber product quality assurance.