Bayesian online changepoint detection-based variable-threshold phase unwrapping algorithm for heterodyne Φ -OTDR

A variable-threshold phase unwrapping algorithm (BOCD-VT) is proposed for the phase-sensitive optical time-domain reflectometer (Φ-OTDR). This algorithm is based on Bayesian theory to online detect changepoints in the phase caused by unwrapping errors. A dynamic unwrapping threshold is set by locally differencing the original signal to constrain the phase unwrapping, thus eliminating potential changepoints. The effectiveness of the algorithm is validated through a series of experiments using a piezoelectric transducer (PZT) as the vibration source. This approach has been implemented in an application using a 16 km urban dark fiber. Experimental results demonstrate that BOCD-VT effectively reduces phase changepoints caused by unwrapping errors when compared to conventional methods, yielding high-quality vibration signals. This approach yields a more than 5 dB improvement in signal-to-noise ratio (SNR) compared to previous post-processing methods, while simultaneously reducing processing time to achieve real-time, jump-free phase unwrapping.