Tracking Australian Wildlife With AI And Aerial Robotics

Tracking Australian Wildlife With AI And Aerial Robotics

Drones are becoming increasingly popular for everything from capturing amazing footage, to training birds of prey and saving lives. But now thanks to a Queensland University of Technology (QUT) collaboration , cutting edge hardware and software is being used to help monitor Australian wildlife.

In a test earlier this month, the Australian Zoo Wildlife Hospital supported a test of a project that could be used to assess koala populations with autonomous multi-copters. An aerial robot was flown both manually and autonomously across a section of bushland around the hospital. Of course it’s not just as simple as flying around and grabbing some pictures of the sleepy marsupials. Amber Gillet, Senior Veterinarian at the Australian Zoo Wildlife Hospital says that detecting koalas is a resource intensive task – it can take 3 staff more than 30 minutes to identify 10 koalas amongst the trees, even when they are in a pen. Out in the wild, this means that population counts are often underestimated and can be inaccurate. The solution from QUT is to combine the drones unique vantage with thermal imaging capabilities that can see through the trees and help spot otherwise hard to see koalas.

It’s not enough to have some cool thermal images of koalas either. Project leader Dr Felipe Gonalez from the Australian Research Centre for Aerospace Automation (ARCAA) at QUT says that the team is using artificial intelligence and machine learning to develop and improve their counting and tracking algorithms. Dr Sandra Johnson from ARC Centre of Excellence for Mathematical and Statistical Frontiers (ACEMS) at QUT has said that Bayesian statistical modelling combined with modern simulation software helps tackle challenging real life issues to better predict wildlife populations with uncertain data. In other words, AI and the hard kind of math are helping drones count our furry friends.

The upshot of this project and collaboration so far is a marked improvement in koala detection rates and more accurate population counts. The technology isn’t just koala specific either — it has the potential to improve researcher’s ability to monitor other wildlife such as dingoes, or damaging species such as feral pigs.

It doesn’t just stop at wildlife — QUT is the base of operations for the Australian Research Council Centre of Excellence in Robotic Vision which is working to create a new generation of robots that can visually sense and understand complex and unstructured real world environments. And sure, it does sound a little bit like the beginning of an Aussie Skynet, but at least we know the Koalas will be well looked after.

Want to know more? Head over to the Queensland University of Technology and the Australian Zoo.