SharkSpotter AI app wins national innovation award!
SharkSpotter, the artificially intelligent shark detection system developed at UTS, for shark spotting by drone, is the national AI or Machine Learning Innovation of the Year, presented at the Australian Information industry Association’s (AIIA) annual iAwards.
In a collaboration with industry partner The Ripper Group, SharkSpotter is a world-first software system that allows for faster reaction times to potential shark threats. Westpac Little Ripper has a suite of Unmanned Aerial Vehicles (UAVS, drones) created to react quickly and efficiently to situations at sea where lives are at risk. The drones are loaded with the SharkSpotter AI application which can efficiently distinguish and identify sharks in real-time using image processing techniques, state-of-the-art sensors and software.
SharkSpotter was developed by a team led by Professor Michael Blumenstein working with Dr Nabin Sharma, both from the Centre of Artificial Intelligence in the School of Software, and several Masters/ PhD students and Research Assistants.
Dr Sharma explains how deep learning algorithms and image processing techniques examine live video feeds from the drones at sea to detect the presence of sharks and their potential threat to water users.
This automated system for detection and identification of sharks in particular, and marine life/objects more generally, uses cutting edge deep learning neural networks and image processing techniques for object recognition and classification.
The system has 90 per cent accuracy in detecting sharks, with the ability to distinguish different species and from 16 other categories of marine life including dolphins, rays and whales. It also identifies surfers, swimmers, boats and other objects in the water.
This visual information is relayed immediately to shore for interpretation to emergency services, beach lifeguards, and water users for appropriate decision-making. Information is sent to a control station on the beach where a human responder will have the final say on what action to take – this could be continued monitoring of the shark to see if it moves away from swimmers or, if it appears to become a direct threat, to sound alarms and advise evacuation.
SharkSpotter is a great example of how an AI application assists humans as it has significantly higher rates of visual accuracy in shark detection than we have, at 90 per cent. At this level of accuracy the drone will certainly help us to improve detection rates, and to maintain safer conditions for water users.
SharkSpotter was also a finalist in two other national categories, having already won three NSW iAwards in 2018 (AI or Machine Learning Innovation, Community Service Markets and Research & Development Project).
Professor Blumenstein reiterates the real-world impact of partnership with industry.
The technology developed by our team is a fantastic example of collaboration between University researchers and industry to create an end-to-end software solution which helps our community and is being used to save lives. We are proud of our on-going relationship with the Ripper Group and we continue to look for ways to enhance their R&D capabilities so that they can deploy their next technology solution.
Already deployed at dozens of beaches in Queensland and New South Wales, many more local councils and the surf lifesaving community throughout Australia are interested to deploy SharkSpotter technology to patrol beaches in their jurisdictions.