SAS has joined forces with the UNC Centre for Galapagos Studies (CGS) to apply crowd-driven artificial intelligence (AI) and machine learning to help protect endangered sea turtles – similar to a recent project it was involved in to help track deforestation in the Amazon.

Through an app called ConserVision, citizen scientists are invited to match images of turtles’ facial markings to help train a SAS computer vision model. Once the model can accurately identify turtles individually, researchers will have valuable information more quickly to better track each turtle’s health and migratory patterns over periods of time. The goal is that in future, the model can perform facial recognition on any sea turtle image whether it comes from a conservation group or a vacationing tourist.

“As our challenges as a global community become increasingly more complex, we need dynamic ways to access and use information to ramp up conservation efforts,” says Sarah Hiser, MSc and principal technical architect at SAS. “By using technology like analytics, AI, and machine learning to quantify the natural world we gain knowledge to help protect ecosystems and tackle climate change.”

In addition to turtles, the Galapagos Islands are home to many unique species not found anywhere else on Earth. An ecological haven for researchers since Charles Darwin first set foot there in 1835, they are also home to the Galapagos Science Centre, a research facility jointly run by UNC-Chapel Hill and the Universidad San Francisco de Quito in Ecuador.

“For over 10 years, the Galapagos Science Centre has hosted exceptional scientists doing innovative research that increases our understanding of the environment and results in positive real-world outcomes,” says UNC-Chapel Hill interim vice-chancellor for research, Dr Penny Gordon-Larsen. “This innovative public-private partnership with SAS will enhance the cenrer’s capacity for analysing data that will positively impact both the environment and the people who inhabit these magnificent islands.”

SAS will initially help UNC CGS with three projects focusing on marine life.
• Sea turtle facial recognition. By using computer vision and machine learning, researchers are looking to identify individual turtles as well as create a health index regarding growth rates, health threats, and presence data. Information gleaned from this data set could be used to understand temporal and spatial movement patterns of these turtles and to identify health risks due to marine debris, boat strikes, diseases, etc. By creating a base health assessment for each animal, a temporal scale between photo capture can be used to determine health over time. Turtles can also be grouped based on region and travel behaviour to provide comparisons for a relative health index and location over time, including the support and participation from citizen scientists and outreach.
• Hammerhead shark patterns. Hammerhead sharks commonly make offshore movements at night and further understanding is needed on what drives these movements. The hammerhead shark project focuses on the presence or absence of these sharks offshore and inshore, and if their movement is synchronous and/or rhythmical. Are there influences from ocean currents, food sources, ocean temperatures, or salinity requirements that help direct the movement of these sharks? Insights can help with developing the boundaries of conservation areas as well as sustainable fishing and tourism practices.
• Phytoplankton predictions. As the basic energy and resource foundation of every food web on the planet, understanding the spatial and temporal dispersal and community interactions of phytoplankton populations is imperative for conservation. The goal of the phytoplankton project is to understand the physical, chemical, and biological factors that influence these small but mighty organisms. Changing ocean temperatures, weather patterns and human activity influence the marine environment that hosts them. Being able to predict changes at the bottom of the food chain can provide insights about the impact of climate change on all marine life and, in turn, human populations that depend on them.