Artificial intelligence used to count manatee clusters in real time: FAU researchers

ORLANDO, Fla. - Manatees are an endangered species susceptible to environmental changes. Researchers who track the manatee population in Florida say they face challenges with accurate counting due to their herding behavior and water reflections. 

Florida Atlantic University's College of Engineering and Computer Science has pioneered new artificial intelligence that employs a deep learning-based crowd-counting approach using CCTV camera images to count manatees in real time. 

MORE HEADLINES:

• Melbourne to discuss $45m worth of upgrades to railroad crossings
• Orlando's Exploria Stadium, home of Orlando City soccer, now has a new name
• Why Florida laws around corporal punishment are 'intentionally vague'

The method, which was recently detailed in a study published in Scientific Reports, utilizes this technology to transform images into manatee-customized density maps. The innovative approach outperformed traditional methods, according to FAU researchers, offering a cost-effective solution for real-time manatee counting. 

According to FAU, the research, sponsored by the United States National Science Foundation, provides a promising future for broader applications in counting techniques for convex-shaped objects.

Manatees, considered endangered, are vital to monitor for population dynamics, especially in the face of human-related threats. The study's dataset and source code are accessible on GitHub for further research and validation.