Detection Performance Improvement of Distributed Vibration Sensor Based on Curvelet Denoising Method.

Area of Science:

• Optical Engineering
• Signal Processing
• Sensing Technology

Background:

• Distributed fiber vibration sensing systems using phase-sensitive optical time domain reflectometry (PhOTDR) are susceptible to time-domain noise.
• Noise in raw Rayleigh backscattering traces can obscure subtle amplitude differences caused by external vibrations.
• Improving the signal-to-noise ratio (SNR) is crucial for accurate vibration detection and localization.

Purpose of the Study:

• To propose and evaluate a curvelet denoising method for reducing noise in PhOTDR systems.
• To enhance the detection performance and accuracy of vibration events in distributed fiber sensing.
• To improve the SNR of vibration location information.

Main Methods:

• The raw Rayleigh backscattering traces from the PhOTDR system were treated as grayscale images.
• A curvelet transform was applied to eliminate random noise from the backscattering traces.
• The amplitude difference induced by external vibrations was extracted after denoising.

Main Results:

• The curvelet denoising method effectively reduced time-domain noise in the fiber vibration sensing system.
• Vibration events were detected with a spatial resolution of 10 m along a 4 km single-mode fiber.
• The SNR of location information for 50 Hz and 1 kHz vibrations increased by 7.8 dB and 8.0 dB, respectively, compared to conventional methods.

Conclusions:

• The proposed curvelet denoising method demonstrates remarkable capability in eliminating noise for PhOTDR systems.
• This approach significantly improves the SNR and detection performance for distributed fiber vibration sensing.
• The method offers a viable solution for enhancing the accuracy and reliability of fiber optic sensing systems.