Mitigation of multi-path propagation artefacts in acoustic targets with cepstral adaptive filtering

TL;DR

This work tackles the challenge of multi-path and motion-induced artefacts in passive acoustic sensing by introducing a cepstrogram-based filtering pipeline. It combines cepstral analysis with an adaptive band-stop filter to suppress Lloyd’s Mirror energy while preserving target harmonics, improving spectrogram quality and downstream ship-type classification. Across simulated movement and underwater datasets, the method yields measurable gains in distortion metrics and MCC, though amplitude preservation remains a concern. The proposed approach shows promise for time-delay estimation, target recognition, and multi-path robustness, with potential extensions to multi-sensor configurations.

Abstract

Passive acoustic sensing is a cost-effective solution for monitoring moving targets such as vessels and aircraft, but its performance is hindered by complex propagation effects like multi-path reflections and motion-induced artefacts. Existing filtering techniques do not properly incorporate the characteristics of the environment or account for variability in medium properties, limiting their effectiveness in separating source and reflection components. This paper proposes a method for separating target signals from their reflections in a spectrogram. Temporal filtering is applied to cepstral coefficients using an adaptive band-stop filter, which dynamically adjusts its bandwidth based on the relative intensity of the quefrency components. The method improved the signal-to-noise ratio (SNR), log-spectral distance (LSD), and Itakura-Saito (IS) distance across velocities ranging from 10 to 100 metres per second in aircraft noise with simulated motion. It also enhanced the performance of ship-type classification in underwater tasks by 2.28 and 2.62 Matthews Correlation Coefficient percentage points for the DeepShip and VTUAD v2 datasets, respectively. These results demonstrate the potential of the proposed pipeline to improve acoustic target classification and time-delay estimation in multi-path environments, with future work aimed at amplitude preservation and multi-sensor applications.

Figures (10)

• Figure 1: Reflection of an acoustic source on a bounded medium. $z_{s}$ and $z_{r}$ represent the perpendicular distances from the source and the receiver to the boundary, respectively; $r_{r}$ denotes the horizontal distance between the source and the receiver.
• Figure 2: A block diagram representing the steps of the proposed cepstrogram time-filtering method. The band stop filter operates in the time-quefrency domain, processing each quefrency individually along the time axis.
• Figure 3: A block diagram representing the low-pass adaptive filtering process. The variables in the dashed box indicate the parameters of the filter.
• Figure 4: Comparison of the spectrograms (upper row) and cepstrograms (bottom row) of the synthetic dataset components. First column represents the original sample from the Audio Set dataset, $R(f,t)$ is the version with amplitude attenuation due to uniform linear motion, $A(f,t)$ contains the component obtained from the direct path propagation, and $B(f,t)$ contains both the components propagated by the direct and reflected paths.
• Figure 5: Block diagram illustrating the use of the raw and filtered audios for spectrogram classification. The top row illustrates the classification pipeline using the unfiltered spectrogram, the middle row illustrates the use of filtering alone, and the bottom row illustrates the parallel classification using two instances of the model.