Wavelet denoising method of threshold compression for LFM-DAS

Distribute acoustic sensors (DAS) based on linear frequency modulation (LFM) offer high sensitivity and strong resistance to interference, enabling continuous monitoring of physical changes along optical fiber links. However, long-range LFM-based distributed fiber-optic vibration sensing systems are often affected by environmental fluctuations and intrinsic noise from system components, which degrades signal demodulation performance. To address this issue, we propose a soft-threshold denoising algorithm based on Morlet wavelet decomposition and noise standard deviation estimation, implemented on top of hardware filtering. The algorithm effectively suppresses background disturbances and high-frequency noise. Simulation results show that, under a signal-to-noise ratio (SNR) of 3dB, the proposed method achieves excellent noise reduction performance. Experimental validation further demonstrates significant enhancement of the backscattered signal quality: at sensing distances of 20km and 220km, the envelope SNR of vibration-bearing backscattered signals improve from 6.17dB and 2.30dB to 49.63dB and 43.86dB, respectively. These results verify that the proposed approach can substantially improve the signal fidelity and robustness of long-distance LFM- DAS systems under low-SNR conditions.