Parks Victoria deploys AI wildlife recognition tool to process field camera images in hours instead of weeks

Parks Victoria, along with data scientist Peter van Lunteren, developed an AI species recognition model trained on five million images provided by 18 Victorian research and conservation organisations. The system can identify more than 200 native and feral species at greater than 95 per cent accuracy and processes 20 images per second.

The tool eliminates lengthy backlogs from motion-triggered field cameras that capture millions of images annually. Dr Erin Nash confirmed detection of red deer threats in burnt areas within hours—work that previously would have taken weeks—allowing faster response to feral animal control and endangered species protection.

Parks Victoria has released the AI model as open source on its website, allowing users to retain ownership of their data and camera images rather than ceding control to external tech companies. Zoos Victoria is already using the system with attached mini-computers to send text alerts when predators enter protected areas.