A groundbreaking pilot has confirmed potential for real-time detection of sea urchins using AI and underwater robotics, paving the way for smarter reef health monitoring and marine pest management.
The Queensland University of Technology (QUT), with support from the Fisheries Research and Development Corporation (FRDC), has completed an initial research project exploring the use of Artificial Intelligence (AI) and robotics to address the challenge of sea urchin barrens along Australia’s southeast coast.
The longspined sea urchin (Centrostephanus rodgersii) has become a significant threat to reef ecosystems and valuable fisheries, including abalone and rock lobster, due to its overgrazing of kelp and other marine vegetation.
In many areas, this has resulted in extensive “urchin barrens” – bare, unproductive seafloor that compromises biodiversity and economic sustainability.
The project has successfully demonstrated the ability of an AI-powered detection system to accurately identify and count sea urchins in real time.
Using image data collected by divers off the coast of Mallacoota, Victoria, the QUT team trained and tested a robust detection and tracking model that achieved an 85–100 per cent count match to human counts, depending on visibility and image quality and at least three–seven times faster.
“This is a big step forward,” said Professor Matthew Dunbabin, lead researcher on the project.
“We have confirmed it’s feasible to use AI to support the real-time robust detection and tracking of sea urchins.
“Our next step is to trial the technology in the field using QUT’s RangerBot autonomous underwater vehicle.”
The AI system was tested on a computing platform equivalent to that used in
RangerBot and demonstrated real-time processing speeds to enable live tracking and counting of sea urchins during underwater surveys.
RangerBot has already proven its value in marine management, particularly through its successful deployment to detect outbreaks of crown-of-thorns starfish and deploying coral larvae on the Great Barrier Reef – highlighting its versatility and potential in managing other reef threats such as sea urchin overpopulation.
The project, which was intended as a proof-of-concept, demonstrates that the technology could now move on to full implementation and in-water trials using RangerBot in sea conditions along the southeast coast of Australia.
The goal of these more comprehensive trials would be to enhance reef monitoring, map barren areas, and ultimately assess whether AI and robotics can also play a role in urchin population control.
Commercial diver Reiner Hurst of Wilderness Wildcatch, who initiated the project after years of witnessing the growing urchin problem firsthand, believes the technology could be a game changer:
“As a diver, you see the devastation happening under the surface – rich kelp forests reduced to lifeless barrens.
I knew we needed better tools to really understand the scale of the issue and take action.
This project is about giving us better eyes in the water and paving the way for smarter, faster, and more targeted responses.”
Dr Nicola Watts of C4 Impact, who helped facilitate the original research proposal, emphasised the importance of partnerships and place-based innovation:
“This project shows the power of place-based initiatives that bring together researchers, industry, communities and investors to tackle complex challenges,” Dr Watts said.
“When we combine local knowledge with cutting-edge technology and shared intent, we open up new pathways for both environmental restoration and regional economic opportunity.”
The project was delivered by QUT researchers Professor Matthew Dunbabin and Serena Reimers in partnership with Reiner Hurst of Mallacoota-based Wilderness Wildcatch and a team of local divers. The project was supported by the Fisheries Research and Development Corporation.
The Fisheries Research and Development Corporation (FRDC) co-invests with industry and government to deliver research and innovation that supports sustainable, productive and profitable Australian fishing and aquaculture. QUT’s Faculty of Engineering is recognised for its expertise in robotics and AI, exemplified by Professor Matthew Dunbabin’s pioneering work in field robotics, autonomous marine vehicles and environmental monitoring technologies.
