A Denoising Preprocessing Framework via Orthogonal Multi-Tap Null-Steering Beamformer Bank: Facilitating Target...

Area of Science:

• Bioacoustics
• Environmental Monitoring
• Signal Processing

Background:

• Acoustic indices are vital for biodiversity assessment via passive acoustic monitoring.
• Anthropogenic noise in urban-rural areas significantly degrades acoustic index robustness.
• Directional noise sources are a primary challenge in these soundscapes.

Purpose of the Study:

• To develop a denoising preprocessing algorithm for robust acoustic index calculation.
• To enhance the performance of acoustic indices in human-dominated soundscapes.
• To address the limitations of existing methods in the presence of directional interference.

Main Methods:

• Utilized a two-microphone array with a differential beamformer for adaptive null-steering.
• Implemented a parallel bank of orthogonal null-steering beamformers to target interference bands.
• Applied a signal compensation mechanism to mitigate self-cancellation effects.
• Evaluated the method using the frequency-dependent acoustic diversity index.

Main Results:

• The proposed algorithm effectively denoises acoustic recordings in urban-rural soundscapes.
• Achieved high-fidelity acoustic information for index calculation across a wider signal-to-interference-plus-noise ratio (SINR) range.
• Demonstrated robustness against directional anthropogenic interference.

Conclusions:

• The novel preprocessing method significantly improves the reliability of acoustic indices.
• This approach enables more accurate biodiversity assessment in challenging anthropogenic environments.
• The technique offers a robust solution for passive acoustic monitoring in human activity areas.