In a groundbreaking conservation and research initiative, Intersect Australia, a not-for-profit node of the ARDC Nectar Research Cloud, has partnered with the Vertebrate Pest Research Unit (VPRU) of the New South Wales Department of Primary Industries (DPI) to develop an AI-powered image detection system for trapping feral pigs. This cutting-edge technology is designed to automate pest animal control efforts, reducing environmental and agricultural damage across NSW.
Feral pigs and other invasive species pose a significant threat to NSW, causing widespread environmental degradation, disease spread and substantial agricultural losses. To understand the extent of this, DPI operates an extensive network of over 1,000 camera traps across the state, capturing valuable data on local wildlife. DPI then tags and enriches these images with information about the animals observed.
To enhance conservation efforts, DPI collaborated with Intersect Australia to develop an artificial intelligence (AI) enabled camera trap. The AI system, housed in a compact camera, is trained to recognise feral pigs in real-time. Once detected, the camera can trigger a mechanism to close a gate and trap the animal, streamlining the control process without requiring as many human resources.
Glen Charlton, Lead Data Scientist of Intersect Australia’s Advanced Analytics and AI team, explained that the project leverages the Graphical Processing Unit (GPU) Reservation System on the ARDC Nectar Research Cloud (Nectar) to accelerate research impact.
“Collaborating with the DPI, Intersect has trained and developed an object detection model capable of identifying the target species based on the existing collection of images,” Glen said.
Glen explained that the Intersect team started with an open-source algorithm and then used transfer learning to develop an algorithm specific for detecting feral pigs. Nectar’s GPUs enabled the development and training of an AI model to be deployed into bespoke trail cameras that can make decisions in real-time, out in the field.
Dr Paul Meek is a Senior Research Scientist with DPI’s Vertebrate Pest Research Unit and has over 30 years of experience as a wildlife ecologist.
“Our collaboration with Intersect and the use of Nectar’s GPUs enables us to continue to develop technology solutions that enhance research and enable practical management solutions for optimising pest animal management,” Paul said.
The prototype was field tested in July 2025. The prototype was able to make near real-time onboard decisions, operate the gate automatically and alert the research team remotely. Therefore, the prototype proved a concept that could significantly reduce the human resources needed for feral animal management, in particular disease outbreaks like foot-and-mouth disease, and improve the efficiency of conservation efforts.
Beyond this project, Intersect Australia is also working to broaden access to AI and machine learning (ML) tools for researchers nationwide. In collaboration with the ARDC, Intersect’s Advanced Analytics and AI (3AI) team has developed an AI-ready application to make it much easier for Australian researchers to use Nectar GPUs for AI/ML.
This application includes a series of standard Nectar AI/ML virtual machine images, which can enable researchers to utilise Nectar’s GPUs as well as Nectar’s CPU-only virtual machines. These images contain pre-installed, commonly used AI/ML tools, which makes it easier for researchers to use Nectar’s virtual machines and GPUs for their own AI/ML applications. Glen explained the new tools reduce the need for extensive coding and IT knowledge, making AI more accessible to researchers across various disciplines.
“Previously, using Nectar’s GPUs required advanced knowledge of backend operating systems and command line interfaces, which was a barrier for some researchers,” Glen said. “With these new AI-ready images, researchers will have easier access to leverage AI/ML applications on Nectar.”
This advancement will allow more researchers to process large datasets, conduct machine learning and deploy AI/ML models more efficiently. It includes tools for AI/ML development, data operations and ML operations, which enables researchers to utilise best-in-class open-source tools to empower their research. Through Nectar’s Application Catalogue, researchers can now select the type of modelling components they require, such as computer vision tools or large language models, and apply them to their own projects with ease.
“Our aim was to make the infrastructure easier for researchers to apply AI/ML to their own research,” Glen said.
Learn more about AI and ML tools for research.
The ARDC is enabled by the National Collaborative Research Infrastructure Strategy (NCRIS) to support national digital research infrastructure for Australian researchers.
Written by Alysha Huxley, Scientell. Reviewed by Jo Savill and Dr Paul Coddington, ARDC.
Also view an article about this project from Intersect.
