Brillouin frequency shift extraction method based on deep learning and mathematical fitting

Brillouin frequency shift extraction depends on curve fitting of discrete data points, a traditional fitting mode mainly comprises mathematical analysis fitting and deep learning neural network fitting, mathematical analysis fitting often excessively depends on data, and when the signal-to-noise ratio of signals is too low, wrong fitting is often generated; the method purely depends on a deep learning neural network fitting mode, the fitting precision is mainly determined by the number of training sets and the setting of model parameters, the fitting accuracy is difficult to guarantee, and all requirements in the actual situation cannot be met. The fitting method comprehensively utilizing the mathematical analysis fitting mode and the deep learning fitting mode is provided, the advantages of the two methods are comprehensively utilized, on one hand, optimization is provided for the two fitting modes, on the other hand, fitting curves obtained through the two optimized fitting methods are further processed, th