Passive stochastic matched filter for Antarctic blue whale call detection

Area of Science:

• Marine bioacoustics
• Signal processing
• Cetacean research

Background:

• Passive acoustic monitoring is crucial for Antarctic blue whale (ABW) population studies.
• Traditional methods struggle with low signal-to-noise ratio (SNR) and complex ocean noise.
• Accurate detection is vital for understanding ABW behavior and distribution.

Purpose of the Study:

• To propose and evaluate a novel stochastic matched filter (SMF) for enhanced ABW detection.
• To improve the signal-to-noise ratio (SNR) of faint ABW vocalizations.
• To assess the SMF's performance against existing detection methods in realistic ocean conditions.

Main Methods:

• Development of an SMF based on signal second-order statistics and simulated sea noise.
• Offline filter bank computation and online noise/SNR estimation using time-frequency analysis.
• Cross-correlation of SMF output with a combined SMF+MF reference for detection.

Main Results:

• The SMF+MF method significantly decreases false alarm rates while maintaining high detection rates for low SNR ABW calls compared to the matched filter (MF).
• The SMF+MF extends detection range by approximately 30 km compared to the Z-detector under ship noise conditions, with a similar false discovery rate.
• The proposed method demonstrates superior performance in challenging acoustic environments with varying soundscapes.

Conclusions:

• The stochastic matched filter (SMF) offers a significant advancement for passive acoustic monitoring of Antarctic blue whales (ABW).
• This method enhances the detection capabilities in low SNR conditions and complex noise environments, crucial for effective population assessment.
• The SMF provides a more robust and extended range for detecting ABW vocalizations, aiding conservation efforts.