Research On Loss Optimization Based On Digital Image Processing Technology

With the development of optical fiber communication,the use of ribbon optical fiber with the advantages of large capacity and short distance communication has become very common,and the market size of the optical fiber ribbon is becoming larger and larger.As a result,the ribbon optical fiber fusion splicer has also become more and more popular Applications.Optical communication is based on optical power to support,however,when the light conduction in the optical fiber,there will be loss of attenuation of optical power,including the loss of fiber optic transmission and fiber transmission loss.If the attenuation loss exceeds a certain threshold,it will cause poor communication.In this thesis,the optimization of the splice loss and the optimization of the loss of the ribbon fiber splicer are completed.The former innovation point is to start from the imaging system,by optimizing the brightness uniformity of the optical display area to indirectly optimize the welding loss,the use of binary,Huff transform linear detection and other digital image processing technology to achieve the original manual adjustment CMOS offset improvements automated tuning,and experiments to verify that automated tuning satisfies backlight requirements more than manual tuning,automated tuning also provides convenience for operators and reduces complexity of work to align CMOS offset degree,improve the working efficiency of the operator.The latter innovative point is the introduction of the concept of curvature,the study of the welding point of the bending loss contribution to the welding,achieved through the analysis of the weld image to calculate the fusion point of the relevant curvature information and the use of linear regression method,by building The regression model was constructed,and the linear regression equation of estimated loss was constructed by the correlation and the significance analysis and the F test on the feature values.The comparison between the estimated loss of the original algorithm and the estimated loss after the algorithm optimization was made The square error sum of the residuals of the estimated loss and the actual loss is reduced from 0.6141 to 0.0041,which verifies the optimization results of the estimated loss algorithm.