Information Extraction Of Brillouin Distributed Optical Fiber Sensor Based On Cross-correlation Convolution Method

With the rapid development of the optical communication industry,distributed optical fiber sensing system based on Brillouin scattering have received more and more attention in the field of structural detection and temperature detection,and are widely used in various industrial fields such as the national economy and national defense.With the increase in sensing distance and sensing resolution,the amount of data processing in the sensing system has increased rapidly,and traditional data processing methods have become increasingly difficult to meet the real-time requirements of the sensing system.Scholars Researchers have proposed cross-correlation convolution method(XCM),artificial neural network and other methods to meet the real-time requirements of the sensing system,but these methods still have a lot of room for improvement.Starting from the study to meet the accuracy and real-time requirements of the sensing system,this paper studies the correlation-based method and proposes two optimization schemes for its shortcomings.The specific work and innovations are as follows:(1)The existing sparse constraint cross-correlation iterative algorithm may have errors in recovering the Brillouin gain line shape.This paper proposes an optimization algorithm based on sparse constraint cross-correlation method.By optimizing the convolution scheme and constraint function,pre-processing the swept data,two groups of data points with the same distribution on the frequency are obtained for convolution operation,and the original convolution curve line shape is restored correctly,thereby improving the accuracy of the extracted Brillouin frequency shift(BFS).Experimental results show that the BFS extraction error is reduced by about 3 MHz when the SNR is1 d B compared to the sparse constrained cross-correlation iterative algorithm;when the SNR is 20 d B,the BFS extraction error is reduced by about 0.3 MHz.(2)During the research based on the sparse constraint cross-correlation method,it was found that only selecting data points near the peak can significantly reduce the BFS extraction error.An optimization algorithm based on the correlation-based method is proposed to filter out the influence of noisy data and enhance the role of ideal data,so as to achieve the purpose of improving the calculation accuracy of the correlation-based method while the time complexity is basically unchanged.The experimental results show that when the SNR is 5 d B,the BFS extraction error is 1 MHz,which is about 4MHz lower than the correlation-based method,and about 2 MHz lower than theoptimization algorithm based on the sparse constraint cross correlation method.