Automated detection of Antarctic blue whale calls

Abstract

This paper addresses the problem of automated detection of Z-calls emitted by Antarctic blue whales (B. m. intermedia). The proposed solution is based on a subspace detector of sigmoidal-frequency signals with unknown time-varying amplitude. This detection strategy takes into account frequency variations of blue whale calls as well as the presence of other transient sounds that can interfere with Z-calls (such as airguns or other whale calls). The proposed method has been tested on more than 105 h of acoustic data containing about 2200 Z-calls (as found by an experienced human operator). This method is shown to have a correct-detection rate of up to more than 15% better than the extensible bioacoustic tool package, a spectrogram-based correlation detector commonly used to study blue whales. Because the proposed method relies on subspace detection, it does not suffer from some drawbacks of correlation-based detectors. In particular, it does not require the choice of an a priori fixed and subjective template. The analytic expression of the detection performance is also derived, which provides crucial information for higher level analyses such as animal density estimation from acoustic data. Finally, the detection threshold automatically adapts to the soundscape in order not to violate a user-specified false alarm rate.

Seasonal and Diel Vocalization Patterns of Antarctic Blue Whale (Balaenoptera musculus intermedia) in the Southern Indian Ocean: A Multi-Year and Multi-Site Study

“…For such a large amount of acoustic data, we resorted to an automatic Z-detector based on a subspace-detection algorithm [28]. The main advantage of this detector is that it does not suffer from the inherent limitations of the classical correlation-based detectors.…”

“…airguns, other baleen whale calls, seismic events, etc). Yet, the number of false detections generated by such interferences are limited due to the Z-detector characteristics [28]. Nevertheless we develop a method for removing potential false detections.…”

Antarctic Blue Whale calls detection based on an improved version of the stochastic matched filter

“…The signal's reference [5] is updated from spectrogram-analyzed OBSh data. A new method for background noise estimation, based on the observation's medianfiltered STFT, bypasses the selection by an operator.…”

“…As for other blue whales [1], [2], [3], Passive Acoustic Monitoring (PAM) seems to be a very efficient tool to develop automatic Antarctic Blue Whale calls (Z-calls) detection algorithms and then facilitate large data analysis to monitor the specie in vast areas. They are mostly based on signal cross-correlation theory: matched filters have been applied [4], as well as spectrogram-based template matching correlation [3], or more recently subspace-detection algorithm [5], [6]. However, those methods do not perform well at low Signal to Noise Ratio (SNR), whether it is due to high background noise environment (e.g.…”

Detection strategy for long-term acoustic monitoring of blue whale stereotyped and non-stereotyped calls in the Southern Indian Ocean

“…Before detection, the audio files are whitened using a FIR filter whose time-varying impulse response is derived from the noise power-spectral-density estimated every 300 seconds as described in [24,Appendix A]. This process makes the detection insensitive to the different types of background noise encountered in the dataset.…”