Optical Fiber Cable Surface Defect Recognition System Based On Machine Vision And FPGA Implementation

As major operators implement fiber-to-the-home strategies,fiber optics is the core device of communication that connects the work and life of each of us.At present,with the rapid development of communication,the quality requirements for optical fiber are also increasing,and the quality of the external protective cover of the optical cable affects the reliability of the optical fiber.The quality parameters of the external protective cover have become the competition for the optical cable market by major optical cable manufacturers.One of the core.In addition to the selection of the protective layer,the quality of the protective layer of the cable is more caused by various defects in the production process,the protective layer does not meet the requirements.At present,there are various problems in the production line of optical cable.In the production process of optical cable,a series of problems such as bulging,scratching,and shrinkage are prone to occur,so that the life of the optical fiber is not guaranteed and the quality of optical fiber communication is affected.Due to the immature technology,most of the cable manufacturers use the detection method by manual testing,a lot of manpower has been invested,but it is still impossible to improve the detection speed and the production speed,and the false detection rate and the missed detection rate are high.In recent years,with the development of machine vision technology and FPGA(Field Programmable Gate Array),the use of machine vision for the detection of industrial products has been applied in a variety of industrial production.This thesis develops from machine vision theory and FPGA image processing theory,and combines the advantages of both to carry out research work on fiber-optic cable defect recognition based on FPGA.Firstly,focusing on the nature of cable image,the image denoising and enhancement algorithms are studied.According to the noise characteristics of the cable image,a modified detail algorithm based on the improved Laplacian pyramid image is proposed.This algorithm also enhances the details of the cable defect while suppressing the image noise of the cable.Secondly,the image threshold segmentation technology based on OTSU opening and closing reconstruction is improved,and the defects of the fiber optic cable are accurately extracted.Then,the expansion and corrosion algorithms are used to remove the false artifacts and preserve the true optical cable defects.Finally,the sub-module of optical cable defect detection algorithm is optimized and improved.The core technology of each algorithm module transplantation is introduced in detail,and the fiber-optic cable defect detection system based on FPGA is completed.Finally,through the modelsim test,a single camera processing rate of 100 frames per second is achieved,and the system has a maximum processing rate of 300 frames per second.