Real-Time AI Shark Detection Boosting Beach Safety

Introduction

The SharkEye system, developed by the Benioff Ocean Science Laboratory (BOSL) at the University of California, Santa Barbara, is an AI-powered shark detection system that identifies sharks near shorelines in real time.

How SharkEye Works

Equipped with high-resolution cameras, drones capture video footage at California beaches, which is then analyzed by machine learning algorithms to determine the presence of sharks. Text alerts are sent to public safety officials, lifeguards, and the community for immediate action.

Benefits of SharkEye

The real-time detection of sharks enhances the safety of surfers and swimmers along the coastline, especially in areas like Padaro Beach near Santa Barbara where juvenile great white sharks and surfers frequent.

Development and Training

BOSL Project Scientist Neil Nathan mentioned that the computer vision model was trained using NVIDIA T4 GPUs with over 15,000 images from drone surveys at Padaro Beach over 5 years. The model achieved 92% mean average precision after 20 hours of training.

Accuracy and Dashboard

The algorithm can detect sharks even a few feet below the surface, potentially surpassing human ability, especially in challenging visibility conditions. The SharkEye dashboard, in partnership with California State University, Long Beach, displays survey results and ongoing acoustic survey detection.

Future Plans

The SharkEye project could aid marine biologists in studying shark behavior and migration patterns, contributing to conservation and public safety efforts. The BOSL team aims to make SharkEye publicly available for broader use in enhancing beach safety.

To learn more about SharkEye, visit the official website. Check out the detailed story on CNN for additional information.