The previous year was marred by no less than 26 unprovoked attacks on humans by sharks in Australia. Two of the attacks ended up in casualties. The scientists in Australia are working on developing a drone that is able to detect the presence of sharks via a machine learning algorithm. The aim is to use this drone to reduce the number of attacks by sharks on humans.
The research work is being carried out by The Little Ripper Group working in collaboration with researchers from the University of Technology Sydney’s School of Software. The Little Ripper Group is a private entity focusing on developing search and rescue drones. According to the arrangement between the two partners, the Little Ripper will develop the drones while the UTS will provide the algorithm.
According to the Head of the School of Software in the Faculty of Engineering and IT, Professor Michael Blumenstein, the autonomous systems that recognizes and identifies sharks and other marine life has been developed using sophisticated neural networks and image processing approaches.
The image processing system makes use of the real time video feeds to detect sharks. The researchers from UTS are able to detect sharks with 90% precision using an object detection system known as Region-based Convolution Neural Network (RCNN). The system is able to filter the sharks from multitudes of images of other marine life and humans in the water.
Little Ripper has constituted a fleet of drones which have been developed using American and German technologies and provide sharp visuals. The Vapor 55 drone provides a flight time of an hour and is equipped with a gyro-stabilized digital camera with FLIR thermal imaging and optical telephoto zoom.
Eddie Bennet, Little Ripper Lifesacer’s CEO believes that recreational activities on the beach will become a lot safer after the development of these drones and the number of shark attacks will exponentially decrease. The drones will be available for shipment in New South Wales and Queensland this September.