Optical Fiber Image Processing And Docking System Based On FPGA

Due to its unique advantages of broadband,low loss and high speed,optical fiber communication technology has become the main transmission means of modern communication network such as backbone network and core network.However,due to the unique physical properties of the optical fiber,the inner glass fiber is protected by the outer envelope layer.Although it is not easy to be damaged,it can not be recovered as quickly as the cable in maintenance and replacement;And there are many cable lines in the fiber core,which will directly cause the difficulty of fusion.Therefore,the accuracy of optical fiber docking is of great engineering significance for the stability of optical fiber communication and subsequent maintenance or replacement.In this paper,after full investigation and experiment on the structure of optical fiber and the imaging of optical fiber under different camera conditions,a more optimized optical fiber processing image and corresponding data are designed by taking FPGA as the core control unit,which can cooperate with the self-made hardware to realize the high-precision docking of the fiber core.The main work of this paper is as follows1.Build the hardware platform of the optical fiber docking system,based on the domestic conventional optical fiber welding equipment,through the top-down design mode,the whole system is divided into five parts,which are optical fiber imaging module,image processing module,image display module,core docking module and data storage module.The working principle and circuit timing of each module are analyzed,and the driver interface of each module is designed in FPGA to complete the code writing of VGA display device driver,camera driver and SDRAM memory driver,so as to realize the effective data transmission of each part.2.A variety of image processing algorithms such as median filter and Sobel operator are implemented in FPGA.By comparing the advantages and disadvantages of various image algorithms,the best threshold algorithm and morphology iteration times are selected to obtain the best binary image.The optical fiber image read in the imaging module is processed into binary data which is easy to calculate and store through image processing operations such as gray transformation,threshold binarization,filtering,morphological operation and edge extraction,and then the coordinate values of feature points are obtained through the data in the memory.3.The image data in memory is analyzed by Hough transform to get the best line fitting.The display position of the optical fiber on the display is obtained by the best fitting line,and the docking algorithm is designed according to the coordinate distribution.Finally,with the help of zynq core board and opencv library,the optical fiber docking parameters in the image are analyzed.The hardware part of the whole system uses FPGA as the core to control other hardware modules to complete docking function,and the software part uses Verilog to realize functions including but not limited to digital signal processing and image processing algorithm,and completes signal debugging through signaltap platform of vivado.