Intel is using AI to gather real-time data on coral reef health

Project CORaiL monitors prosthetic reefs with AI-powered video cameras.


There’s no question that coral reefs are rapidly deteriorating due to overfishing, bottom trawling, warming temperatures and unsustainable coastal development. In order to help protect and restore reefs, scientists need more data, and Intel thinks it might have a solution. Today, Intel and its partners -- Accenture and the Sulubaaï Environmental Foundation -- announced CORaiL, an AI-powered project to monitor reefs and analyze their resiliency.

In May 2019, the team installed concrete structures, or Sulu-Reef Prosthesis, meant to support unstable coral fragments in the reef surround Pangatalan Island in the Philippines. The structures incorporate fragments of living coral that will grow and expand to provide marine habitat. The team also placed underwater video cameras with Accenture’s AI-powered video analytics platform (Video Analytics Services Platform, VASP) , which counts and classifies marine life, sending real-time data to researchers.

The Accenture platform -- powered by Intel’s Xeon processors and other hardware -- allows researchers to collect vital data without actually being in the water. Divers typically collect video footage manually, but that presents a few drawbacks, like the potential that they’ll interfere with wildlife behavior. In those cases, they also have to view and analyze the data after their dives.

A diver installs a CORaiL AI-powered video camera.

“The value of your data depends on how quickly you can glean insights to make decisions from it,” said Accenture’s Chief Analytics Officer Athina Kanioura. “With the ability to do real-time analysis on streaming video, VASP enables us to tap into a rich data source — in effect doing ‘hands on’ monitoring without disrupting the underwater environment.”

In the last year, the CORaiL team has collected about 40,000 images. Researchers are already using those to assess reef health, and engineers are working on a next-generation CORaiL prototype, with an optimized convolutional neural network and backup power supply. They may add infrared cameras to capture nighttime video as well. In the future, CORaiL might also be used to study the migration rate of tropical fish to colder waters and catch anyone violating restrictions meant to protect the reefs.