09.13.17

Object Recognition Drones for Shark Detection

Worried that sharks might be lurking near your beach? There’s no need to get a bigger boat; instead contact Australia-based Westpac Group, which is offering Shark Spotter, a system that uses drones equipped with object recognition to detect signs of sharks in the water.

Shark Spotter works with Little Ripper drones, which are military-grade unmanned aerial vehicles (UAVs) that cost up to $250,000 each. Starting in September 2017, the drones equipped with Shark Spotter are set to begin patrols along some Australian beaches.

Westpac Shark Spotter was launched a year ago. The organization sponsored a team at the University of Technology (UTS) in Sydney to develop the algorithm and implement it on the rescue drones.

The team developed the algorithm based on a sophisticated deep-learning framework that serves as the backbone for the shark detection and recognition system. The framework employs a Region based-Convolution Neural Network (RCNN) for object detection and object recognition.

Using footage previously collected by the Little Ripper Drones, the university researchers tested the algorithm by having it observe various objects near beaches, including swimmers, surfboards and sharks.

Based on this data, a deep learning system then can conduct real-time object recognition from the video feeds from the UAVs.

The researchers initially have employed a central processing unit to conduct its shark investigations. However, the group eventually plans to adopt a graphics processing unit, which is more suited for running deep-learning algorithms.

At first, the system will conduct shark detection in real time using video downloaded from the UAVs. However, in the future Shark Spotter will detect sharks on the UAVs themselves, only alerting the drone operators when a shark is sighted.

Little Ripper now has 35 drones in its fleet which range in flight capacity from 15 minutes to four hours, according to Digital Trends. The fleet consists of various types of drones, including the Vapor 55 unmanned helicopter from US-based Pulse Aerospace. The fleet also employs the Vario Benzen Unmanned Helicopter from Germany’s Vario Helicopters.

The technology could help reduce incidents of shark attacks on humans. While relatively rare, there were still 84 unprovoked shark attacks on humans in 2016, according to the Florida Museum of Natural History at the University of Florida.

Shark detection is just one application in the fast-growing market for drones in the public safety segment. The global market for public-safety drones is expected to rise to $1.15 billion in 2022, according to the market intelligence firm Research and Markets.

Tyler Schulze is vice president, strategy & development at Veritone.  He serves as general manager for developer partnerships, cognitive engine ecosystem, and media ingestion for the Veritone platform. Learn more about our platform and join the Veritone developer ecosystem today.