Shark attacks are fortunately extremely rare, but when they do occur the results for those involved are usually devastating. The Little Ripper shark spotting drone from Australia aims to change all that by keeping an eye out for these scary denizens of the deep while you swim.
Although rare, when a shark attack happens it’s big news, and for good reason. These apex predators of the deep have the ability to tap into our most primordial fears. Fortunately, various strategies are employed by lifesavers and municipalities around our coastline to minimise the risk of fatal encounters and make us feel safer. Shark nets positioned just behind the surf zone in KZN is one example, as are the dedicated shark spotters located on high ground around False Bay.
However, both of these approaches have their problems. First up, thanks to human predation, many shark species are extremely endangered. And when you remove these apex predators from the food chain, there’s a very real chance the entire ecosystem can collapse. This is the main problem with shark nets. And, although sharks are easy to spot from altitude, there’s no way an observer hanging out on a mountainside with a pair of binoculars can locate them all.
This is where the decidedly high-tech Little Ripper shark spotting drone comes in. Under development by the Australian firm Ripper Group in conjunction with the University of Technology Sydney (UTS). What it does is employ sophisticated artificial intelligence to identify sharks from the air via its camera and transmit the live video feed to lifesavers on the beach, who can then warn swimmers via its onboard loudspeaker when they need to get out of the water fast. The drone also has the ability to drop shark repellent kits into the water, along with other lifesaving aids. And, should an attack occur, rescue personnel will immediately know when and where – giving them the ability to render assistance as soon as possible.
To train the drone’s algorithms to differentiate between sharks, dolphins, surfers, swimmers, waves, rocks and the like, researchers at UTS sourced drone footage from publicly available platforms such as YouTube to manually teach it what common objects look like. So far, a CPU (central processing unit) based system has been used for the research, but the plan is to soon move on to a GPU (graphics processing unit) based system as it’s more optimised for Deep Learning.
Ripper Group’s aim is to initially conduct shark detection in real time via the downlinked video from the drones. Once they have all that completely ironed out they’ll move on to shark detection in real time on the drones themselves – with only detections downlinked to the UAV pilots and operators, making the entire process autonomous. – (c) 2017 NavWorld
Source: Little Ripper
You can check out some great drone footage of sharks swimming amongst surfers below: