Abstract
Technologies for observation of the dangerous production facilities are an important direction. Many practical monitoring problems are solved by using fiber-optic temperature sensors. These sensors are superior to electronic sensors in many aspects – weight, dimensions, and the effect of electromagnetic hindrances. But the most important advantage is the ability to use a single sensor element (fiber) to carry out measurements over several kilometers, which is fundamentally impossible by the electronic measuring instruments. However, fiber-optic measurements are inherently related to the processing of signals with a high level of noise that occurs during optoelectronic conversion. Because of observation systems usually indicated the changes in the measured value, it is necessary to develop methods for identifying small shifts in the signal with the extremely small signal-to-noise ratio. The paper presents a technique for signal processing that characterizes the Stokes and Anti-Stokes components when the fiber is heated in several areas. The technique is based on an algebraic rising in the divergence of the values of the Stokes and Anti-Stokes components with increasing temperature. It was possible to identify the leap of the signal level reliably despite of the increasing the noise spectral density, and it was confirmed by physical modeling.