Australian researchers launch AI koala 'facial recognition'
June 3, 2021
Artificial intelligence (AI) researchers in Australia are experimenting with facial recognition technology to monitor koalas as part of conservation efforts.
The team from Griffith University in the northeast state of Queensland is attempting to use AI to recognize the marsupials when they use wildlife crossings, designed to offer the animals a safe route either over or under public roads.
They hope the data would give a better understanding of how koalas are using them, and whether the crossings could be better located to protect them from getting injured by cars.
"The goal of this project is to set up an AI-based monitoring facility to monitor the koalas' road crossing behaviors, so that we can analyze how many koalas are using the facilities to cross the road using underground pathways or the above-road crossings," Jun Zhou, an associate professor at the university, said on Tuesday.
Zhou is leading the two-year pilot study funded by the Queensland government through a AUD $90,000 ($70,000, €57,000) koala grant scheme.
AI could end the need for manual camera checks or use of identification tags and GPS to identify which animal species were using the crossings, Zhou said.
"Now, with artificial intelligence developing very quickly over the past 10 years, the technology is powerful enough to help recognize not only koalas generally, but which individual koalas are using the crossings," he said.
Up until now, video monitoring of the crossings needed to be checked by humans to see what types of animals had been caught on camera.
Collaborating with conservation groups
The pilot study, "Predicting Koala Road Crossing Behaviours using AI-Powered Observation Network," is expected to be rolled out at koala-friendly crossing locations near the state capital Brisbane with 20 cameras by the end of July 2021.
The researchers plan to work with conservation groups on training AI to differentiate between individual animals based on their appearance and movements.
"Animal movement will trigger image capture, with images transferred to a server at Griffith University, " Zhou said. "Computer vision and machine learning systems will be used to process images, allowing for automatic detection and recognition of individual koalas," he added.
Dwindling numbers amid bushfires and road collisions
The furry mammals live along Australia's eastern coast, typically spending most of their days sleeping in tree branches. But increasing habitat loss due to logging, bushfires, construction and expanding development is driving the species towards extinction. Though not classified as endangered, Australian authorities do consider the species "vulnerable," and the International Union for the Conservation of Nature (IUCN) classifies koalas as "threatened" on its "Red List" of endangered species.
From 1997-2018, an average of 356 koalas entered care facilities due to vehicle collisions each year.
Griffith University said preventing koala fatalities and injuries caused by vehicles was one of the most important tasks for koala conservation.
Following the bushfires of 2019-2020 — which claimed the lives of an estimated 5,000 koalas — a New South Wales parliamentary inquiry found that koalas would be extinct in the state by 2050 without urgent intervention to protect their habitat.