Initiative
Applying artificial intelligence for flora and fauna research using image recognition technology.
Overview
A proof of concept is underway using a Microsoft Cognitive Services computer vision system to analyse pictures from camera traps to identify target species for review by scientists.
Previously scientists had to manually review all images from camera traps, including many triggered by the wind or unwanted species.
The next stage in the project is refining and training the system to recognise the taxonomy of the animals captured in the images.
Opportunities are also being examined to use Cognitive Services to analyse drone footage recorded by CSIRO scientists in Kakadu National Park to detect Salvinia, an introduced weed that infests water bodies.
Benefits
The use of these advanced digital technologies will allow scientists to collect and analyse higher volumes of environmental data for research and development of more effective environmental management policies.
2020 progress update
Scientists and researchers are extending the use of image recognition and other artificial intelligence services to extract valuable data from ground-based and aerial photography.