2023_programme: Bayesian method for underwater acoustic data processing in an uncertain environment
- Session: 21. Underwater Noise - Modelling and Measurements
Organiser(s): Alexander Gavrilov
- Lecture: Bayesian method for underwater acoustic data processing in an uncertain environment
Paper ID: 2049
Author(s): Pihan-Le Bars Hélène, Kinda Bazile, Perrier de le Bathie Cécile
Presenter: Pihan-Le Bars Helene
Abstract: Underwater acoustics has become an important means of monitoring the state of marine ecosystems in recent years, through the sound produced by animals or the quantification of anthropogenic acoustic pressures. However, acoustic monitoring in general requires the deployment of permanent or long-term networks. Such monitoring generates large amounts of acoustic data requiring the development of automated processing algorithms, either to detect signals of interest, or to classify and identify target species.\nThe estimation of noise and its counterpart in detection are common issues in signal processing. The complexity of the marine environment makes this development task challenging. On the one hand, the acoustic data are site-dependent due to the physical properties of the environment that shapes the acoustic propagation, and the spatio-temporal distribution of marine animals. On the other hand, the background noise is modulated by weather conditions and human activity in the vicinity of the acoustic recording sites. \nAn automatic processing algorithm, not requiring any prior knowledge of the marine environment, has been developed. This algorithm, adapted to large volumes of acoustic data, is based on a Bayesian and recursive approach of joint estimation of the background noise and the probability of presence of any signal in the time-frequency domain. \nBased on this solution, we present several applications for ambient noise monitoring and rough soundscape characterization including the separation of near field (identifiable vessels) and far field (ambient noise) noise contributors, and a study of environmental conditions impact in the Bay of Biscay.
- Corresponding author: Dr Hélène Pihan-Le Bars
Affiliation: Shom & French Naval Academy Research Institute
Country: France
e-mail:
