Artificial Intelligence Improves Real-Time Manatee Counts

A recent study published in Scientific Reports suggests that artificial intelligence can be instrumental in accurately and affordably counting manatees in real-time. The study conducted by Florida Atlantic University (FAU) utilizes “deep learning” to teach a computer to recognize the unique elliptical shape of manatees in surveillance camera images. This development could help biologists react more swiftly to threats faced by the species.

Counting manatees by hand in complex outdoor environments presents several technical challenges, including ripples, glare, overlapping manatees, and observer biases. By leveraging deep learning-based crowd counting approaches, an AI system can automatically count manatees in designated regions. This technique has been conventionally used for counting human crowds in urban planning and public safety applications.

Xingquan (Hill) Zhu, the lead researcher and professor in FAU’s Department of Electrical Engineering and Computer Science, states that this AI approach provides a better understanding of manatee populations. Monitoring manatees with limited resources has been a challenge, and traditional methods have fallen short due to factors like warm weather and the ongoing COVID-19 pandemic.

Accurate manatee population estimates are vital for protecting the species and managing resources effectively. The study comes amid alarming trends in manatee deaths, with a record number of deaths reported in 2021. By automating the counting process and eliminating errors caused by perspective distortions and overlapping shapes, the AI system developed by FAU outperforms traditional methods.

The researchers used readily available video surveillance camera footage to estimate manatee density and calculate their numbers. By applying an Anisotropic Gaussian Kernel technique, the AI system was taught to recognize the elliptical shape of manatees, reducing overlap errors common in other counting methods. The next step for the researchers is to improve the accuracy of AI counting further.

State wildlife biologists may be able to utilize this AI-based method for future manatee counts, using existing surveillance cameras in state parks. The study was sponsored by the United States National Science Foundation, highlighting the potential of AI in wildlife conservation and management.

The source of the article is from the blog meltyfan.es

Privacy policy
Contact