Environmental company Astron Environmental Services confirms this week that it’s developing a first in intelligent shark mitigation methods alongside local tech start-up Airbotix Pty Ltd.
The system aptly named Sentinel VDS will leverage drone technology to capture live video footage, with highly sophisticated detection software working autonomously to warn beach users if target shark species are detected.
Astron is utilising its expanding geospatial capabilities for the development of this intelligent real-time shark detection software, and has suitably partnered with Airbotix who are leading the build and design of the aerial hardware.
Julian Kruger, Technical Director of Astron said, “As a surfer and an environmentalist I understand the importance and need that we have to implement a shark attack mitigation approach which doesn’t unnecessarily cull sharks and kill other by-catch species.”
“All of the research being conducted tells us that shark attacks are on the rise due to increasing numbers of people entering the water. For this reason it’s essential that we learn to patrol our beaches in a way that ensures user safety, but not at a cost to the environment.
“If we can provide ocean users with an effective early warning system as a means of mitigating against shark attacks in a non-invasive and non-lethal manner, then we see this as a win for both people and sharks.
“The system we’re developing is all about building on current methods of visual detection such as beach spotters and helicopter patrols. But because we are adopting a lower impact and lower cost platform we are able to achieve a much higher degree of surveillance for any particular site.
“The fact that we are also using smart optics and computer vision technology to considerably enhance the rate of detection means that much of the human error can be avoided.”
Software development for the real-time shark detection system will be led by Astron Scientist Dr Mehdi Ravanbakhsh.
Dr Ravanbakhsh is a remote sensing expert whose research includes the use of underwater stereo-video technology for measurement and automation of fishery assay and sampling.
“To date we have been trialing a number of machine learning and computer vision approaches to detect sharks as anomalies and moving objects in the ocean,” explains Dr Ravanbakhsh.
“We are now looking to move beyond detection to more sophisticated analysis of these objects, which will allow us to increase the rate of identification for sharks.”
The system is being developed to function in most conditions and along most coastlines.
“The system will prove ideal in most conditions and being a method which doesn’t have a permanent fixture it can be easily deployed to new locations, including remote areas,” said Mr Kruger.
“We see that this system could offer an invaluable safety tool for some of our iconic beach events such as the Margaret River Pro, the Rottnest Swim and our Surf Life Saving carnivals.
“Being a flexible and low cost option there is also massive potential for it to be adopted more broadly, particularly for those activities where our ocean use occurs in discrete locations, like surfing.”
Astron has been working on computer vision processing algorithms to automatically detect sharks in near shore environments since last year.
“Whilst we’re still in the early phases of development, we are extremely excited and encouraged by our results to date,” explains Mr Kruger.
The next phase for Astron involves the collection of additional data to train the software to improve rates of detection and identification.
“In parallel to the software that we’re developing, our partner Airbotix is undertaking the hardware design, build, integration and testing of the aerial platform which is based on an already established design.”
“We expect to be undertaking fully functional aerial trials with the Sentinel VDS throughout the second half of this year,” said Mr Kruger.
We're interested to talk to potential partners, so please give us a call or contact us through our contact page. Continue to follow our exciting developments via the Sentinel VDS Facebook or Instagram accounts.