The Special Interest Group for Underwater Acoustics (SIGUA) is holding a 2-day residential event at Guyers House, near Bath on 27-28 January 2025 to explore solutions to research challenges set by industry. This will be the third workshop, following two previous events in 2022 (Sheffield) and 2023 (Bath).
We will also be holding a virtual pre-event on Zoom from 14:00 to 16:30 on 10 December 2024. This will enable Challenge Leaders to introduce their datasets and provide early access, and for participants to get to know one another and organise into teams.
This is an opportunity to be exposed to some of the current and future challenges in underwater acoustics and to work together with other researchers, to explore and develop new/innovative solutions. This will be a ‘hackathon’ style event, with small teams working on one of the challenges together for most of the two days, while representatives from industry move between the groups to provide application context and to answer questions.
Overnight stays will include breakfast, lunch and dinner. Day attendees will be provided with lunch. A minibus transfer to/from Bath Spa train station will also be provided.
Challenge #1 – Thales: Detection, Tracking and Classification for Passive Sonar
Sonar operators are faced with the task of detecting, tracking and classifying contacts in an environment with an extremely high prevalence of noise and clutter. Operating in this highly complex environment induces false alarms and missed detections that must be managed in order to provide sufficient tracking capabilities. In this challenge, you will be tasked with developing an algorithm to assist the sonar operator to perform their role. Provided with a passive underwater acoustic data set, you will be asked to track and classify, using tonals, objects of interest in the water. You should visualise your output in a form appropriate for presentation to the sonar operator. Your algorithm must handle the many features of underwater acoustic data that make the role of the sonar operator so difficult. These features include ambient noise, crossing contacts, quiet contacts, clutter from sea mammals or other vessels in the water and the impact of water characteristics on the propagation of sound. You have the freedom to focus on one or more aspect of sonar operation to create an algorithm that will ease the burden of the sonar operator’s role.
Click here for further challenge details
Challenge #2 – ORE Catapult & Celtic Sea Power: Automatic Characterisation of Harbour and Grey Seal Vocalisations
Collaborative efforts within the Offshore Renewable Energy Industry have focused on reducing consenting times, enhancing coexistence, and minimising environmental impacts. A key strategy to achieve this involves expanding knowledge on the presence and behavior of less-studied marine species, such as grey and harbour seals, which are particularly abundant in the UK. To support this, pre-labeled datasets of grey and harbour seal vocalisations – both in the wild and captivity – will be made available, along with limited video footage of the sites. The primary goal of this challenge is to improve the performance of automatic detection and classification algorithms for seal vocalisations, enabling the development of more advanced methods that can not only distinguish between species but also detect social behaviors.
Click here for further challenge details
Challenge #3 – University of Southampton / STEM-CCS: Measurement of CO2 Leaks from the Seabed
STEMM-CCS (Strategies for Environmental Monitoring of Marine Carbon Capture and Storage) was an EU funded project with the goal of demonstrating the effectiveness of different monitoring of marine CCS facilities. The objective is to detect CO2 bubbles which, in the unlikely event of a leak occurring, would indicate the release of CO2 in gaseous form. Passive acoustic data were collected on an array of 5 hydrophones during a controlled release experiment in the North Sea, in which CO2 was released 3 m into the sediment to simulate a leak. The acoustic array recorded the sounds of the bubbles made as they entered the water column to estimate the gas flux (the volume of leaking gas per unit time) and determine the locations of the leaks. The various noise sources corrupting the data make this task challenging, including ambient ocean noise, high-frequency noise from other seafloor instruments deployed during the experiment, and noise from nearby ships and from distant seismic surveys. Participants must develop innovative approaches to analysing this noisy data to improve the detection and quantification of CO2 leaks and localise their sources accurately. Success in this challenge could lead to more reliable monitoring of marine CCS sites, helping to ensure their environmental integrity.
Click here for further challenge details
Posted on 10th October 2024 in Events, Underwater Acoustics
