September is coupled with the arrival of spring, which inspires a flurry of activity in the animal world as breeding season for many different species cranks up a notch.
September is also Save The Koala month, which aims to highlight the threats to koalas, including habitat loss, climate change, disease and human activity, contributing to population declines.
It’s for these reasons that the innovative work from a team of Griffith University AI researchers and ecologists has once again been funded to develop ‘facial recognition’ camera technology at koala crossing locations across South East Queensland.
This project, now in its third year, determines how and which koalas are using the crossings and ultimately provide research-based planning to help protect the declining population.
Professor Jun Zhou, from Griffith’s School of Information and Communication Technology, leads the pilot study, which was initially funded by a $90,000 Community Sustainability Action Grant awarded to the team by the Queensland Government’s Department of Environment and Science (DES) in March 2021.
The work was subsequently funded again with a $100,000 grant from DES in June 2023.
“This project extends our innovative AI-powered koala monitoring system to cover wider areas of koala habitat in South East Queensland, and engage with 14 local community groups across 10 local government areas to facilitate the installation and maintenance of the camera network,” Professor Zhou said.
In July 2021, the team rolled out 24 AI-powered cameras at koala crossing locations in the Redland City Council area.
The cameras were automatically triggered by koala movement and the hundreds of captured videos and images were sent to a server at Griffith University.
The team then develop the AI to automatically analyse the videos and images to recognise the individual koalas.
Caroline Moss, Queensland Rail Group Senior Manager Environment & Sustainability, said: “A research project like this helps us to understand how this technology can be applied, not only here in the Redlands, but where appropriate in other locations, given that Queensland Rail operates a really large network.”
As the AI needs to recognise individual koalas using the crossing at monitoring sites, the research team worked closely with conservation groups such as Koala Action Group Redlands, Daisy Hill Koala Centre, Moggill Koala Rehabilitation Centre and Currumbin Wildlife Sanctuary to train the technology to distinguish one koala to another based on their appearance and movements.
“We can see from the data that koalas are in trouble here,” project co-researcher Dr Douglas Kerlin said.
“They’ve recently been listed as endangered in Australia, and Southeast Queensland was formally a hotspot. The driver of that decline has been urbanisation.”
There is a reported link between koala mortality that is caused by cars and breeding and population dynamics of koalas.
During the breeding season, Frank Mikula from Lone Pine Koala Sanctuary said his team sees an increase in the number of koalas coming into the sanctuary’s care.
“We’re all about sharing information and it’s really good to be able to put people in contact and create that conduit between researchers so that everyone can get to the same goal faster,” Mikula said.
“The technology does the heavy lifting for us, and that’s really important moving forward.
“With increased knowledge about how koalas are crossing roads, we can better inform mitigation and management so that we can ensure a better long-term future for koalas.”