The Wavelet Filtration Denoising in the Raman Distributed Temperature Sensing

Abstract

Up to now a Distributed optical fiber Temperature Sensor (DTS) based on the Raman scattering exhibits a relatively low characteristic causes sharp temperature changes to be improved. Modeling metrological characteristics has a long history but essential progress did not achieved. This paper presents a novel technique of the extremal filtration developed to improve the DTS temperature and partially spatial resolution. The algorithm is based on wavelet transformation of backscattered anti-Stokes and Stokes signals and deconvolution on high-frequency components with total regularization of variations. Experimental results correctly agree with real modeling values with denoising ones. The advantage is the ability to reconstruct temperature with 0.01 degree accuracy at fixed spatial resolution. We present a lot of simulations and figures demonstrated the efficacy of the proposed technique.