Home Artificial Intelligence facial recognition could help save the species

facial recognition could help save the species

by Joey De Leon

The koala population in northeastern Australia is endangered, and one of the biggest challenges in preserving their numbers is accurately counting them. However, researchers from Griffith University are using a combination of community groups, camera systems, and artificial intelligence (AI) to address this problem.

Koalas are notoriously difficult to spot and monitor due to their elusive nature and unpredictable movement patterns. This makes it challenging to determine which conservation efforts are effective in helping the species thrive. Traditional methods of manual counting or tagging and monitoring have proven insufficient.

To overcome these limitations, Professor Jun Zhou and his team at Griffith University embarked on a pilot study using cameras and AI technology. They placed 24 cameras at designated “koala crossing locations” in the Brisbane and Redland city council areas. These cameras were triggered by koala movement, automatically recording any activity.

The researchers then enlisted the help of community koala monitoring groups to train an AI system using the gathered images. The AI was trained not only to identify koalas but also to recognize individual animals. This innovative approach provided a tool that could be used to monitor koala populations in much larger areas.

By deploying 100 cameras in a one square kilometer research region, the AI-powered system could continuously monitor the entire area, 24/7. It could analyze video recordings to identify and count koala movements, including the ability to recognize new individuals in the region.

While the results of the pilot study have revealed positive outcomes in terms of technology efficacy, there have been concerning estimates regarding the decline of koala numbers. Urbanization has been a significant driver of this decline in the Southeast Queensland region.

To further improve the accuracy of the technology, the Queensland Government has granted additional funding for the expansion of the trial to 10 local government areas. Professor Zhou hopes that more local councils will adopt the technology in the future. However, he emphasizes the need for community involvement in deploying and maintaining the camera network, as this task cannot be fully automated.

In conclusion, the combination of camera systems and AI technology offers a promising solution to the challenge of counting and monitoring koala populations. This technology can provide invaluable data for conservation efforts and help identify strategies to protect and preserve this endangered species.

You may also like