Improving the measurement accuracy of distributed RH sensor by using adaptive 2D bilateral processing method

Abstract

The measurement accuracy of the distributed relative humidity (RH) sensor based on optical frequency domain reflectometry (OFDR) is constrained by the number of data points involved in the cross-correlation calculation. Reducing the number of data points to achieve high spatial resolution measurements can significantly compromise the accuracy of the results. This effect is especially evident in the presence of small strains, such as those induced by humidity variations. To overcome this problem, this study proposes and experimentally demonstrates a distributed RH sensor with high accuracy by using polyimide-overlaid fiber and image processing technology. Cross-correlation data arrays are generated by processing the collected distributed Rayleigh scattering signals along the sensing fiber, allowing the construction of a two-dimensional (2D) image that captures spectrum shifts due to humidity-induced strain (HIS). Thus, the adaptive 2D bilateral processing method is introduced to realize distributed RH measurement with high spatial resolution. Experimental results show that the abnormal information can be removed and the real information can be extracted by the proposed method. The relative humidity gradient information can be reconstructed free of outliers under spatial resolution of 4 mm with a wavelength sweeping range of 20 nm over a 48-meter test fiber. Additionally, the peak-to-peak measurement errors from the proposed technique are half of those from traditional processing methods and more precise humidity-induced strain information can be reflected. This straightforward sensing technique paves the way for the development of cost-effective, reliable and practical distributed humidity sensing systems.