Shark hunting AI-powered algorithm to begin flying over Aussie beaches

A SHARK hunting algorithm which uses live video footage streamed from drones to detect sharks will be rolled out on Aussie beaches next month.

SHARK wary surfers are about to get some peace of mind with a new artificially intelligent drone monitoring system set to begin patrolling Australian beaches next month.

The project is a collaboration between the Westpac and Telstra backed Little Ripper Group and a team of researchers at the University of Technology Sydney who developed an algorithm capable of using real time video footage streamed from drones to detect sharks and alert swimmers.

Professor Michael Blumenstein is the head of the School of Software in the Faculty of Engineering and Information Technology at the university. He and his team have been working on the “world first shark spotting” algorithm for the past year.

In developing and refining the algorithm, the more data the better, he said. Researchers used known migration patterns, recorded sightings and first-hand drone footage filmed over popular beaches to train the SharkSpotter system to detect sharks swimming beneath the surface.

At the heart of SharkSpotter is a machine learning technology known as Deep Learning which is part of a broader family of machine learning methods based on learning data representations.

“At the end of the day, AI is going to be at the centre of everything. It’s going to be the fuel that drives things forward,” Prof Blumenstein told news.com.au.

The SharkSpotter drones are due to be deployed across some of Australia’s most popular beaches next month in the hope the system can be used to save the lives and limbs of Aussie swimmers and surfers.

According to its creators, the algorithm is about 90 per cent accurate in its ability to distinguish sharks from dolphins, whales and other marine life — which is a massive increase in accuracy when compared to the human eye from planes or helicopters above.

For conservationists, it could also be the best current solution that allows sharks and surfers to coexist in the water as previous deployments of protective nets have been criticised by environmentalists for their potential to harm marine wildlife.

“That’s one of the reasons we’re so excited about it,” Prof Blumenstein said. “You don’t want to be intrusive in that environment.”

Prof Michael Blumenstein lead a team at UTS who developed the shark hunting algorithm.

When it spots a shark, the drone will hover over the animal and can warn swimmers through a megaphone.

According to Dr Paul Scully-Power, a co-founder of the Little Ripper Group an emergency beacon and a life raft can even be dropped from the drone to any distressed surfers below.

The drone system allows for “a rapid response that can be done with less people,” said Prof Blumenstein. And when it comes to cost, while the drone technology, maintenance and training for staff runs into the tens of thousands of dollars, “you can really scale that up and save government and the community (money).”

There are currently 35 drones in the fleet which range in flight capacity from 15 minutes to four hours, according to the Little Ripper Group.

With a recent spate of shark attacks along Australia’s coast, such initiatives are no doubt welcome by authorities.

Australia ranked behind only the United States in the number of unprovoked shark encounters with humans last year, according to the International Shark Attack File of the University of Florida.