Processing of FBG sensor signal by improved wavelet

Demodulation system is the core sensing technology of optical fiber Bragg grating (FBG). De-noising and seeking the peak wavelength are the two important factors that influence the demodulation accuracy. In order to obtain the precise wavelength drift, aiming at the defects of the traditional demodulation methods, a new wavelength detection scheme for optical fiber Bragg grating (FBG) was put forward, which was composed of an improved de-noising method and the Gaussian fitting peak searching algorithm. The translation invariant wavelet combined with the new threshold value and the proposed new threshold function method to deal with noisy FBG sensor signal; then the Gaussian fitting peak searching algorithm was adopted to find the peak wavelength of the de-noised signal for further analysis. The experiment result shows that the improved invariant wavelet can deal with noisy FBG sensor signal with different SNR and get a higher SNR value and a lower mean square error value than the other wavelet de-noising methods mentioned; this kind of wavelength detection technique can get the measurement maximum peak error of less than 1 pm, which means a much higher accuracy than general wavelength detection methods.