Real-time noise elimination method oriented to POTDR opposite scattered light signals

Abstract

The invention discloses a real-time noise elimination method POTDR opposite scattered light signals. The method comprises the following steps: firstly carrying out Fourier transform of fractional order on collected data frames to obtain fractional domain data; carrying out noise elimination and reconstruction on an imaginary part of the fractional domain data via a wavelet transformation method; carrying out signal reconstruction by using the processed imaginary part and a real part of the fractional domain data; and carrying out Fourier transform of fractional order on the signals obtained by reconstruction and obtaining noise-eliminate signals according to transform results. Through the real-time noise elimination method, the signals are processed due to combination of fractional order Fourier noise elimination and wavelet noise elimination; the background noise in the opposite Rayleigh scattered light can be eliminated in real time; meanwhile, the effective signals can be saved to a maximum extent; the signal-to-noise ratio of sensing signals can be increased; compared with wavelet noise elimination and average noise elimination, the problem that the useful peak signals in the signals are leveled because of the wavelet noise elimination can be solved; meanwhile, the problems of long consumed time and low real-time performance of the average noise elimination can also be solved.

Description

technical field [0001] The invention belongs to the technical field of optical fiber sensing, and more specifically relates to a real-time denoising method for POTDR backscattered light signals. Background technique [0002] Distributed optical fiber sensors are widely used in the measurement of temperature distribution and pressure and stress distribution in a large spatial range due to their advantages of single-ended, non-destructive measurement and omniscience. The POTDR (Polarization Optical Time Domain Reflectometry) system has unique advantages in the fields of vibration frequency measurement and intrusive sensing, but the backscattering signal it collects is extremely weak and the signal-to-noise ratio is low. The sensing signal is inevitably noisy, especially at the tail of the signal curve. Due to the large attenuation of optical power, the signal is more seriously affected by noise, which inevitably complicates the signal and is not conducive to subsequent analysi...

AI Technical Summary

Problems solved by technology

The POTDR (Polarization Optical Time Domain Reflectometry) system has unique advantages in the fields of vibration frequency measurement and intrusive sensing, but the backscattering signal it collects is extremely weak and the signal-to-noise ratio is low. The sensing signal is inevitably noisy, especially at the tail of the signal curve. Due to the large attenuation of optical power, the signal is more seriously affected by noise, which inevitably complicates the signal and is not conducive to subsequent analysis and signal processing.

[0003] The method of adopting average processing can effectively reduce random noise, but can not meet real-time requirements; the public number is the wavelet denoising algorithm disclosed in the Chinese patent application document of CN105157874A, although it can meet the requirements of real-time denoising, but its The signal is smoothed according to the characteristics of the mother wavelet, so that the effective peak signals in the actual signal are filtered out together, so that the effect of improving the signal-to-noise ratio after denoising is not ideal, especially not suitable for POTDR systems;

[0004] The Fractional Fourier Transform (FRFT) method can transform the time-domain signal to any time-frequency axis and is not limited to the frequency domain. The signal whose energy is not concentrated in the frequency domain may be concentrated in a certain fractional domain. The effect of filtering noise; however, the fractional Fourier transform still transforms the entire time domain, and like the Fourier transform, it cannot analyze the local characteristics of the signal, and the local characteristics of the signal are very important in the optical fiber sensing system;

[0005] For the denoising processing of the optical fiber sensing signal, the processing method based on average denoising or wavelet denoising may affect the accuracy of the analysis and judgment of the signal in the later stage; while the FRFT method cannot analyze the local characteristics of the signal

Images