Research On Image Processing Of Brillouin Scattering Spectrum In BOTDA System

In the development trend of industrial informatization,distributed optical fiber sensing system has the advantages of small size,high pressure resistance,corrosion resistance,anti-interference and low cost.With its natural and unique advantages,it has attracted the attention of a large number of researchers.Brillouin Optical Time Domain Analysis(BOTDA)is one of technologies.It can measure the temperature and strain of the sensing fiber by monitoring parameter changes.And it is currently used in aerospace,rail transit,disaster warning and safety monitoring in bridge construction and other fields widely.How to improve the performance indicators of the BOTDA system has become a research hotspot at home and abroad.First of all,this article starts from the principle of Brillouin scattering effect and analyzes the sensing mechanism of the BOTDA system based on the Stimulated Brillouin Scattering(SBS)effect.The three-wave coupling equation of the SBS process is calculated and the Brillouin scattering spectrum is simulated.Then the performance indicators of the BOTDA system are briefly introduced,after that the system is built and the key components are analyzed.The system of temperature measurement experiments is completed.Then,because of the image characteristics for Brillouin scatter spectrum,this paper uses image processing filtering methods such as Gaussian filtering and median filtering to denoise the Brillouin scatter spectrum.Gaussian filtering could attenuate highfrequency components that exceed the effective signal effectively,however,the lines and other fine image details may be lost when targeting large noisy images.While median filtering can eliminate more noise while retaining edge details.Therefore,hybrid filtering can be used on the Brillouin scattering spectrum,and median filtering can be used before Gaussian filtering to reduce the noise amplitude and avoid the loss of image edge details.Through the calculations,Gaussian filtering increases the image signal-to-noise ratio by 3.99 d B,median filtering increases the image signal-to-noise ratio by 6.67 d B,and hybrid filtering increases the image signal-to-noise ratio by 9.18 d B.It is known that hybrid filtering has the best performance among the above three methods.Finally,this paper studies the methods of extracting the Brillouin frequency shift for the BOTDA system.The amplitudes of the Brillouin scattering spectrum are regard as the pixels of the image.In view of the complexity for the curve fitting calculation and the long processing time,we propose to use the peak position method to quickly extract the Brillouin frequency shift without increasing the system complexity and sacrificing spatial resolution.Compared with the traditional centroid analysis method in image processing,the processing speed is increased and the accuracy of extracting frequency shift is improved.In addition,this paper also proposes an image edge detection calls Canny to extract Brillouin frequency shift.In the field of image processing,the edge is the junction where the gray level changes obviously on the image.In the Brillouin scattering spectrum,the step change is also similar to the gray level change in the image.According to the principle of Canny,the image will be filtered first,then the gradient amplitude and direction are calculated,after that nonmaximum value suppression and double threshold detection are performed to extract the Brillouin frequency shift.Finally,the effectiveness of Canny in the extraction of Brillouin frequency shift is verified through experiments.The image processing method has better overall characteristics compared with the traditional fitting methods.It can avoid repeated fitting operations when mutations occur at multiple positions,and greatly reduce the data processing steps and processing time.