Characterization on Imaging Properties of Optical Fiber Array Based on Machine Vision

The imaging qualities of optical fiber array were discussed in this paper by characterization with various self-developed instruments based on machine vision. The basic processes include extracting clear image transmitted by optical fiber array, transferring the extracted image to the calculation unit, processing it by digital methods such as gray value and binary system. And then the size, shape and color of defects will be judged according to the distribution, brightness and color of pixels. In combination with the standard for controlling various defects of optical fiber array, the characterization results will be further calculated and judged and finally output. The fiber-optic imaging element defect detection device is to detect and analyze dots, chicken wire and grids. The contrast ratio detection device is to measure the crosstalk among optical fibers within optical fiber array. Meanwhile, the imaging evenness could also be characterized by achieving the transmittance differences between periphery and central zone. The reproductivity reaches above 97%. The relative transmittance tester is to measure the relative transmittance located at any positions and compare the evenness among different locations. The precision reaches microscale. An automatic detector recently developed by our team is to detect scratches on sample surface with measurement accuracy exceeding 99%, solving the great challenge. It is because of application of machine vision that the corresponding characterization methods and equipment are innovated. This promotes the realization of digital, intelligent, automated and high précised characterization.