Using drones and artificial intelligence to monitor large colonies of seabirds can be as effective as traditional methods on the ground, while reducing costs, labor and the risk of human error.
Scientists at Duke University and the Wildlife Conservation Society (WCS) used a deep learning algorithm – a form of artificial intelligence – to analyze more than 10,000 drone images of mixed colonies of seabirds in the Falkland Islands off the Argentine coast.
The Falklands, also known as the Malvinas, are home to the world’s largest colonies of black-browed albatross (Thalassarche melanophris) and the second largest colonies of southern jumping penguins (Eudyptes c. Chrysocome). Hundreds of thousands of birds breed on the islands in densely dispersed groups.
The deep learning algorithm correctly identified and counted albatrosses with 97% accuracy and penguins with 87%. In total, automated counts were less than 5% of human counts about 90% of the time.
“Using drone surveys and deep learning gives us a remarkably accurate, less disruptive and much easier alternative. One person or a small team can do it, and the equipment you need to do it isn’t that expensive or complicated. “said Madeline C. Hayes, remote sensing analyst at Duke University Marine Lab, who led the study.
Monitoring of the colonies, which are located on two rocky and uninhabited outer islands, has so far been carried out by teams of scientists who count the number of each species observed on part of the islands and extrapolate these numbers to obtain estimates of population for all colonies. Since colonies are large and densely dispersed, and penguins are much smaller than albatrosses (and, therefore, easy to miss), counts often need to be repeated. It is a laborious process, and the presence of scientists can disrupt the breeding and parenting behaviors of birds.
To conduct the new investigations, WCS scientists used a consumer drone to collect more than 10,000 individual photos, which Hayes converted into a large-scale composite visual using image processing software.
She then analyzed the image using a convolutional neural network (CNN), a type of AI that uses a deep learning algorithm to analyze an image and differentiate and count the objects in it. “Sees” – in this case two different species of seabirds. These counts were added together to create complete estimates of the total number of birds found in the colonies.
“A CNN is loosely modeled on the human neural network, in that it learns from experience,” said David W. Johnston, director of the Duke Marine Robotics and Remote Sensing Lab. “You train the computer to detect different visual patterns, like those created by black-browed albatrosses or southern jumping penguins in sample images, and over time, it learns to identify the objects forming those patterns in other images such as our composite photo. . “
Johnston, who is also an associate professor of the practice of marine conservation ecology at Duke’s Nicholas School of the Environment, said the emerging drone and CNN-based approach is widely applicable “and greatly increases our ability to monitor the size and health of seabird colonies. in the world and the health of the marine ecosystems they inhabit. “
Guillermo Harris, senior environmentalist at WCS, is a co-author of the study. He said: “Counting large colonies of mixed-species seabirds in remote locations is an ongoing challenge for conservationists. This technology will contribute to regular population assessments of certain species, helping us to better understand whether conservation efforts are working. “
Making and training CNN can seem daunting, Hayes noted, but “there are tons of resources online to help you out, or, if you don’t want to deal with it, you can use a free, pre-built CNN and the personalize. do what you need. With a little patience and guidance, anyone can do it. In fact, the code to recreate our models is available online to help other researchers start their work.