Denoising algorithm of Φ-OTDR systems based on adaptive fractional wavelet transform denoising

Area of Science:

• Photonics and Sensing Technologies
• Signal Processing
• Wavelet Theory

Background:

• Phase-sensitive optical time-domain reflectometry (Φ-OTDR) is crucial for long-range sensing.
• Background noise and low signal-to-noise ratio (SNR) challenge Φ-OTDR performance in complex environments.
• Existing denoising methods struggle to preserve signal fidelity while suppressing broadband noise.

Purpose of the Study:

• To propose a novel adaptive fractional wavelet transform (NFRWT) for denoising Φ-OTDR signals.
• To enhance SNR and suppress background noise in long-range sensing and engineering applications.
• To improve the performance of distributed acoustic sensing (DAS) signal analysis.

Main Methods:

• Constructed a time-domain coupled wavelet basis with fractional-order phase modulation and translation phase compensation.
• Developed an adaptive fractional-order selection strategy maximizing NFRWT-domain energy concentration.
• Applied multiscale wavelet decomposition and soft-threshold shrinkage for noise suppression and waveform preservation.

Main Results:

• The NFRWT method effectively distinguishes useful signal components from broadband background noise.
• Demonstrated superior performance compared to existing denoising algorithms in 20.2 km long-range sensing.
• Achieved an average SNR of 63.38 dB and a noise floor of -84.32 dB rad²/Hz in intrusion monitoring.

Conclusions:

• The proposed NFRWT-based denoising method significantly improves Φ-OTDR signal quality.
• NFRWT offers robust background noise suppression and waveform fidelity preservation.
• The method is effective for long-range sensing and real-world engineering scenarios like intrusion detection.