Prototype

Underwater Noise Identification

Underwater noise identification from acoustic recordings using spectrograms

Aim

To develop, using the iMagine platform, a prototype service for processing acoustic underwater recordings for identification and recognition of marine species and other noise types (e.g., offshore piling).

Development actions during iMagine

1

Setting up development environment at iMagine platform

2

Connecting the Mongo database to the iMagine platform for ingesting and extracting data

3

Developing, training and refining the AI model at the iMagine platform

4

Enhancing the FAIRness of data output in accordance with community standards and relevant vocabularies

5

Documenting approach and resulting prototype

6

Contributing to dissemination and outreach

Expected Results

1

The underwater sound classification service has large potential to support the scientific community in marine biodiversity and ecosystem research. Underwater sound and noise pollution are increasingly recognized as an essential indicator of healthy seas and oceans. 

Detecting and classifying different species can contribute to species abundance assessment, and detection and characterization of noise pollution can provide insight in effects of human activities on marine life.

2

With this, the scientific interest in underwater sound detection and classification is rapidly increasing. Potential policy supporting examples are downstream applications in monitoring for MSFD (as part of the 11th GES) and OSPAR Underwater noise indicators.

Involved Partners

VLIZ