Face is familiar: Researchers learn to tell our koalas apart in order to keep them safe – In Qld

By June 4, 2021 News

INSIGHTS        Katrina Beikoff        Friday        June 04, 2021

 Queensland researchers are launching koala “facial recognition” to save koala lives by monitoring how the animals across south-east Queensland are using man-made wildlife crossings.

The team from Griffith University will use artificial intelligence (AI) to recognise individual koalas when they use the road crossings in a bid to cut the number of koalas being hit by cars.

An AI-powered koala observation network of 20 cameras will be rolled out at koala crossings across the Redland City Council area by the end of July.

The team will work with conservation and wildlife hospital groups such as Koala Action Group, Daisy Hill Koala Centre, Moggill Koala Rehabilitation Centre and Currumbin Wildlife Sanctuary to train the technology to distinguish one koala from another based on their appearance and movements.

Between 1997 and 2018, an average of 356 koalas each year had to be taken to care facilities after being hit and injured by vehicles. Project lead, Associate Professor Jun Zhou from Griffith’s School of Information and Communication Technology, said reducing koala fatalities and injuries caused by vehicles was one of the most important tasks for koala conservation.

Being able to better predict koala road crossing behaviour using the latest technology was vital, he said.

Currently cameras monitor koala crossings, but each of the captured videos then has to be manually checked to see whether the animals filmed using the crossings were koalas or other species.

“Now, with artificial intelligence developing very quickly over the past 10 years, the technology is powerful enough to help recognise not only koalas generally, but which individual koalas are using the crossings using videos that have been trained by our AI,” Zhou said “This last goal is quite challenging, but we hope the research and collaboration with wildlife organisations will make it possible.”

He said the network involved interconnected devices each including a camera, a motion sensor, a wireless/mobile network module and a solar panel.

“Animal movement will trigger image capture, with images transferred to a server at Griffith University. Computer vision and machine learning systems will be used to process images, allowing for automatic detection and recognition of individual koalas,” he said.

This data would then be analysed by the team to provide a greater understanding of koalas’ use of crossing structures and help experts and government choose the best locations and improve the design of fauna crossings.