Bioacoustical Periodic Pulse Train Signal Detection and Classification using Spectrogram Intensity Binarization and Energy Projection

TL;DR

This paper presents a novel multi-stage spectrogram-based method for automatic detection and classification of minke whale pulse train signals, achieving promising accuracy in challenging acoustic environments.

Contribution

It introduces a new detection and recognition approach combining spectrogram edge detection, energy projection, and machine learning for whale song analysis.

Findings

Achieved 63% true positive rate at 0.6% false positive rate.

System maintained 84% precision and 71% F1 score in real-world recordings.

Method effectively handles broadband, narrowband noises, and low SNR conditions.

Abstract

The following work outlines an approach for automatic detection and recognition of periodic pulse train signals using a multi-stage process based on spectrogram edge detection, energy projection and classification. The method has been implemented to automatically detect and recognize pulse train songs of minke whales. While the long term goal of this work is to properly identify and detect minke songs from large multi-year datasets, this effort was developed using sounds off the coast of Massachusetts, in the Stellwagen Bank National Marine Sanctuary. The detection methodology is presented and evaluated on 232 continuous hours of acoustic recordings and a qualitative analysis of machine learning classifiers and their performance is described. The trained automatic detection and classification system is applied to 120 continuous hours, comprised of various challenges such as broadband and narrowband noises, low SNR, and other pulse train signatures. This automatic system achieves a TPR of 63% for FPR of 0.6% (or 0.87 FP/h), at a Precision (PPV) of 84% and an F1 score of 71%.