Mapping Australian wildlife with the help of drones, AI and citizen scientists

Queensland University of Technology is teaming up with Landcare Australia and wildlife rescue organisation WIRES to expand its drone-based koala detection program via an Australia-wide network of citizen scientists.

“This is a really aspirational project, we want to change the way conservation is done,” says QUT Associate Professor Grant Hamilton.

“We aren’t going to do conservation just by publishing papers,” he says.

The partnership ‘WildSeek’ aims to build a national conservation AI network, involving citizen scientists across the country collecting data via drone to be analysed by QUT’s National Conservation AI Analytics hub.

Last week, QUT ran a three-day drone workshop for community and Landcare group members and volunteers, focused on using drones fitted with thermal imaging sensors to capture data on local native animals. 

The initial groups based in Noosa, Queensland, Tamworth, NSW and East Gippsland, Victoria are all working on koala conservation. 

Once collected, drone data will be analysed by QUT’s artificial intelligence koala algorithm. Hamilton says the AI detects a koala’s thermal signature based from a “hot blob of a particular range of temperatures in a particular position”, producing a koala count and map of their locations.

While the project is initially focused on koalas, Hamilton believes the concept could be used for other species as well.

“The WildSeek Project’s initial focus is on identifying koalas, but it has the potential to expand to include multiple species including kangaroos, wallabies, birds and wombats, as well as identify invasive pests or other threats,” he says.

By helping more people on the ground – citizen scientists – to collect conservation data using drones, the program has the potential to broaden the impact of the research to every corner of Australia, Hamilton says.

“Can you imagine just how powerful that would be?” he says.