• Subscribe

Spotting a shark before it spots you

Speed read
  • Australia recorded 26 shark attacks in 2016
  • Video footage used to train deep learning algorithm
  • SharkSpotter can differentiate between a shark, a whale, and a person

Unmanned aerial vehicles (UAVs), also known as drones, are all the rage these days. We’re using them in journalism, filmmaking, agriculture, disaster management, law enforcement, and numerous other fields.

Drones will soon be performing routine tasks like delivering packages — or saving you from a shark attack.

Let 'er rip. Westpac Little Ripper Lifesaver UAVs feed live video to the SharkSpotter deep learning algorithm. With a 90 percent accuracy rate, SharkSpotter differentiates between marine life and ocean vessels, alerting humans to the presence of sharks. Courtesy Little Ripper.

Okay, shark attacks aren’t routine, but Australia recorded twenty-six last year. This is a concern for people who spend time near the ocean.

SharkSpotter, a new technology developed by Michael Blumenstein and Nabin Sharma at the University of Technology Sydney (UTS), is helping swimmers and surfers feel a little safer while enjoying ocean recreation. SharkSpotter uses a deep learning algorithm that mines real time video footage collected by Westpac Little Ripper Lifesaver drones.

Work on SharkSpotter began when Little Ripper approached UTS with a plan to explore shark detection using aerial imagery. The researchers examined live video feeds gathered over time by camera-fitted Little Ripper drones and manually annotated the video footage to indicate the specific location of sharks.

Annotations also identified other marine life such as whales, rays, and dolphins, as well as swimmers, surfers, and boats.

Blumenstein's team then trained the deep learning algorithm for object detection and classification using the video frames and annotations. SharkSpotter can now differentiate between sharks, dolphins, surfers, boats, and other objects in the water with a 90 percent accuracy rate.

SharkSpotter began a trial period over coastal waters in September. When a shark is detected, the system notifies Surf Life Saving personnel via text message. A recorded voice alert is also played from a megaphone attached to the drone or installed on the beach.

Sharma says they are testing these alert methods to find out what works best and doesn’t create panic. <strong>Beach buddies. </strong> Operations manager Ben Trollope, David Wright, mayor of Ballina Shire, and Nabin Sharma (from left) confer about the Little Ripper UAV, a drone that gathers video to identify sharks from the air. The SharkSpotter system uses this video to improve human safety without endangering marine life. Courtesy Little Ripper.

The algorithm will be fine-tuned based on how it performs in the trial. “It is not expected that an artificial intelligence system will work straight away after deployment,” Sharma says, “as there are many unknown scenarios.” So far, he notes, the system is working well.

Who benefits from this technology?

Certainly, with SharkSpotter in operation, swimmers and surfers have a better chance of avoiding a shark attack, but the sharks have something to gain as well.

Protective nets have been deployed in the water to increase public safety, but their use has the potential to harm marine wildlife.

In contrast, Sharma’s drone solution doesn’t interfere with the marine ecosystem. “It is quite obvious that beachgoers and beach recreation will become safer,” Sharma notes, “but it is also important for us not to disturb the marine life.”

Dr. Sharma says shark attacks are not becoming more common, but that as the human population increases, beach recreation is gaining popularity and people are spending more time in the ocean.

To avoid shark attack, Sharma advises beach users to learn about water safety and shark behavior.

But if SharkSpotter works out, swimmers and surfers will have someone looking over their shoulder as they hang ten.

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2021 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.