The key to reversing the decline in Australia's threatned wildlife is taking practical steps to deliver effective on-ground management. Technological innovations are providing exciting opportunities to achieve greater efficiencies and generate substantially better outcomes for biodiversity across the country. AWC is committed to trialling and deploying a suite of innovative new technologies in our science program and we welcome your support for this program.
Camera traps are essential tools for collecting data on species diversity and abundance. Traps are activated by motion sensors and record a colossal number of images: a 3 month survey alone can capture more than 500,000 photos! Around 60% of these images are set off by non-animal motion such as leaves blowing in the wind, making camera traps both a boon and a bane for AWC ecologists, as every image is checked manually. As a result, AWC is harnessing the power of technology to increase efficiency
The AWC AI Program has successfully developed the 'level 1' model. Excitingly, AWC has now built its first Species Classifier (level 2) model. This machine learning technology has been 'trained' to identify common species in a specific area. The next stage is to use footage obtained by thermal cameras mounted on drones to count or classify animals on sanctuaries. Trials are underway at AWC's Faure Island Wildlife Sanctuary off the coast of Western Australia.
Continued improvements to the AI processing will enable AWC to count specific species and identify the existence of targeted feral species such as Feral Cats. This branch of the AI program requires industry leading research into analysing moving footage and attempt identification by thermal signature, or by specific motion type (e.g. 'hopping' vs 'hunting' motion of the animal). We also envisage developing a 'level 3' model that will identify individual animals based on their stripes or patterns
This project has provided additional documentation in a PDF file (projdoc.pdf).
