Distributed High-Temperature Sensing Based on Non-Local Haar Transform in OFDR

Abstract

We demonstrate a short-time long distance distributed high-temperature sensing by non-local Haar transform (NLH) in optical frequency domain reflectometry (OFDR). By searching similar pixels across a non-local region, NLH makes good use of incomplete space similarity of information contained in 3D cross-correlation distribution of local Rayleigh scattering spectra measured by OFDR, which can be used to enhance image denoising performance and retain rare details of original spectra data. With the proposed method, we have achieved a short-time distributed high-temperature sensing ranging from 950°C to 1050°C over 102 m by reduced-cladding single mode fiber (RC-SMF) with a sensing spatial resolution of 2 cm. Compared with traditional image denoising methods including Gaussian filter (GF), block-matching three dimensional filter (BM3D) and wavelet denoising (WD), the proposed NLH method has a best performance to restore the consistency of spectral shift distribution caused by the same temperature change without deterioration of sensing spatial resolution.