Improving the accuracy of measuring the peak frequency of the Mandelstam-Brillouin scattering spectrum using the least squares method

The paper presents a detailed analysis of various noise sources affecting the operation of Brillouin optical analyzers, which allows a deep insight into the fundamental limitations of this type of fiberoptic sensors. The main task of fiber sensors based on the Brillouin scattering principle is to extract the peak frequency of the scattering spectrum, since the peak frequency shift provides information about the temperature and deformation of the fiber. Due to the high noise level, quadratic fit is often used in data processing. Formulas for the dependence of the minimally detectable Brillouin frequency shift on the signal-to-noise ratio and frequency step are given.