Artificial intelligence and the dugong
Better protecting marine megafauna through social media and artificial intelligence is the goal of Laura Mannocci's work in New Caledonia. It's a project that involves taking to the skies... in a microlight aircraft.
Dugongs, rays, sharks, turtles... these large representatives of what is known as marine megafauna are particularly vulnerable. " These animals reproduce late and have few offspring, making them fragile species that require special protection," explains Laura Mannocci, a marine biologist at the Marbec laboratory. Fewer than 1,000 dugongs, these large marine herbivores, are thought to still inhabit the waters of New Caledonia.
To better protect these majestic animals, we must first improve our census data. Researchers at Marbec, Entropie, and Lirmm laboratories have taken on this task, deploying an innovative method that involves the use of artificial intelligence, " tools that ecologists are increasingly adopting," explains Laura Mannocci.
Aerial survey
Where censuses are traditionally conducted below sea level, Laura Mannocci chose to take to the skies. More precisely, she flew her microlight aircraft 50 to 60 meters above sea level, at which altitude she surveyed a portion of New Caledonia's western lagoon. "These waters are very shallow, 2 meters at most, so you can see marine animals even at this altitude,"explains Laura Mannocci. To make sure she didn't miss anything, the researcher attached a camera under the microlight aircraft to capture the ballet of the rays and the passage of the turtles.
Filming is one thing, but identifying is another... And that's where algorithms come in. "We used what is known as deep learning , artificial intelligence applied to pattern recognition," explains the biologist. The researchers developed an "intelligent" program designed to view the images filmed while flying over the lagoon and identify a school of rays here or a dugong hidden next to a coral head there.
Social media
To identify these animals, the recognition algorithm that "learns to learn" must first have seen many of them. "We needed images from multiple angles and in various habitats," explains the biologist. How could we collect so many images of an elusive animal like the shark or a rare animal like the dugong, for example? To overcome this obstacle, scientists turned to social media. "The dugong is such an iconic animal that we figured when people had the chance to film it, they would probably post the images on social media," explains Laura Mannocci.
After searching Facebook, Instagram, and YouTube, scientists identified 20 videos of dugongs, which they used to "feed" the algorithm and teach it to recognize the animal. The result: thanks to deep learning, the program is able to detect and identify 80% ofthe animals in the videos. "This method offers a new and effective way to count and map dugongs and other charismatic marine species with the aim of better protecting them. These censuses could lead to recommendations to place certain areas under strict protection if necessary," explains Laura Mannocci.
