Acoustic characterization of bird vocalizations using distributed fiber optic sensing based on Φ-OTDR

Abstract

Aiming at the drawbacks of conventional wildlife acoustic surveillance such as inadequate spatial coverage, restrictions on equipment installation, and the poor real-time performance of traditional monitoring methods, this paper proposes a bird vocalization signal acquisition and acoustic characterization model that relies on the Phase-Sensitive Optical TimeDomain Reflectometry (Φ-OTDR). The system uses optical fiber as a distributed acoustic sensor to measure avian vocalization signals through the detection of phase changes in the Rayleigh backscattering signals produced by soundwave vibrations in the fiber. Five birds (cuckoo, oriole, sparrow, crow and magpie) were taken as subjects of study, 21 sets of acoustic signal data of undisturbed background and perturbed recordings were gathered. Signal quality improvement was realized by wavelet denoising (coif4, 2-level soft threshold) and high-pass filtering. Time- and frequency-domain indicators such as RMS amplitude, peak amplitude, dominant frequency and bandwidth were compared between species. The findings indicate that various species of birds show considerable differences in time-domain energy and frequency domain spectral structure, which confirms the practicality of the use of Φ-OTDR-based avian acoustic monitoring. This study provides both theoretical support and experimental evidence of distributed fiber optic wildlife acoustic monitoring systems.