Design Of Submarine Cable Defect Recognition System Based On FPGA

Submarine cables are cables that are laid on the ocean floor and are an integral part of today's international communications.Compared with ordinary optical cables,the submarine cable must be designed to ensure that the internal optical fiber can adapt to the seabed pressure,wear,corrosion,biological environment,etc.,and the difficulty of laying the submarine cable,the submarine cable needs to reach a working life of more than 25 years.Defects in the surface of the submarine cable production process,such as small holes,bumps,bulges,bubbles,necking,surface scratches,discoloration,etc.,will have a fatal blow to the quality of the submarine cable,affecting the performance of the submarine communication system and shortening the life..At present,domestic submarine cable manufacturers mainly rely on manual testing,and the detection efficiency is not high.With the continuous development of machine vision technology in recent years,the use of machines instead of manual inspection has become the development trend of industrial modernization.In this paper,related research on surface defect recognition based on FPGA is carried out,and the following contents are completed:(1)Starting from the hardware structure suitable for surface defect detection of submarine cable,the design idea of image processing platform based on FPGA processing chip is discussed,and an image processing platform including storage circuit,video input and output circuit and configuration circuit is constructed.Analysis solves the difficulties of each part.(2)The design of image preprocessing algorithm and background culling algorithm in the submarine cable surface defect recognition algorithm is studied.Among them,the submarine cable image preprocessing algorithm mainly focuses on the image noise due to uneven illumination and the inconspicuous defect points.A detail enhancement denoising algorithm based on image pyramid is proposed.The background rejection is based on the Sobel edge detection algorithm.The cable is partially removed to facilitate better display of internal defects.(3)Using the OTSU image segmentation and the mean-based double histogram equalization algorithm to calibrate the defect area around the background of the cable with the background removed,and use the morphological operation,watershed segmentation and other processes to remove the micro defects and pseudo defects,and finally The defect is extracted from the cable image.(4)The core algorithm steps of the submarine cable defect recognition algorithm are adapted to the hardware migration around the FPGA hardware platform,including hardware implementation of bilateral filtering,Sobel edge detection,OTSU threshold segmentation,morphological operation,etc.The algorithm is used in real-time video to meet the requirements of real-time detection of submarine cable defects.