جامعة الفلوجة

This project presents an AI-powered real-time highway traffic accident detection and emergency response system that combines YOLOv10x object detection, BoT-SORT multi-object tracking, and a novel TimeLocker event stabilization engine to accurately detect and verify traffic accidents while minimizing false alarms. The system automatically dispatches emergency alerts to hospitals and traffic authorities within milliseconds, operates at 30 FPS, achieves 80.5% [email protected] with a 0.77 F1-score, and requires only existing surveillance cameras without additional roadside infrastructure. Designed for deployment on remote highways such as the 400 km Expressway 1 in Al Anbar, Iraq, it provides a scalable, cost-effective solution that significantly reduces emergency response time and enhances road safety.