AI Case Study
University of Queensland researchers reduce amount of time needed to process images of coral reef to assess health using machine learning
Scientists catalogue and classify pictures take of coral reef off Sulawesi Island in Indonesia to assess the health of the reef. The AI system automates classification of the photos which reduces the amount of manual time needed to process them from over 10 minutes to seconds.
Public And Social Sector
Education And Academia
The technology was deployed in a survey of Sulawesi Island's coral reef. "A combination of 360-degree imaging tech and Artificial Intelligence (AI) allowed scientists to gather and analyse more than 56,000 images of shallow water reefs. Over the course of a six-week voyage, the team deployed underwater scooters fitted with 360 degree cameras that allowed them to photograph up to 1.5 miles of reef per dive, covering a total of 1487 square miles in total. Researchers at the University of Queensland in Australia then used cutting edge AI software to handle the normally laborious process of identifying and cataloguing the reef imagery."
"The broad scientific consensus is that coral reef ecosystems worldwide risk collapse by as early as 2050, if CO2 emissions continue at the current rate. This sobering reality has catalysed a range of actions aimed at quickly assessing the health of reef ecosystems worldwide. Indonesia is situated in the heart of the Coral Triangle - home to the greatest levels of marine biodiversity on the planet. There are reefs here that contain more species than the entire Caribbean, which is why the bioregion is of particular interest to scientists looking into reef resilience. Coral reefs provide food security to half a billion people and contribute around US$375 billion per year to the global economy. If they collapse globally, the ocean fishing industry will collapse with them."
The use of deep learning has sped up image processing by automating it: "what would take a coral reef scientist 10 to 15 minutes now takes the machine a few seconds".
"Using the latest Deep Learning tech, they ‘taught’ the AI how to detect patterns in the complex contours and textures of the reef imagery and thus recognise different types of coral and other reef invertebrates. Once the AI had shown between 400 and 600 images, it was able to process images autonomously."
Training was done on 400-600 photographic images of coral.