An AI-Based Real-Time Road Safety Monitoring System for Smart Cities

Due mostly to slow emergency response and delayed hazard detection, traffic accidents continue to rank among the world's leading causes of mortality. Conventional road monitoring systems frequently fall short of offering real-time and intelligent analysis of dynamic traffic conditions since they mainly rely on manual surveillance and static rule-based methods. Vigilant RoadShield, an AI-driven real-time road hazard identification framework developed utilising the YOLOv8 deep learning model combined with a Django-based web monitoring dashboard, is presented in this study in order to overcome these constraints.Using live video or picture inputs, the suggested system automatically identifies road dangers like cars, obstructions, and possible collision situations by utilising computer vision algorithms. For high-speed object identification with optimal accuracy, YOLOv8 Nano is used, guaranteeing low latency performance appropriate for real-time deployment. An interactive online interface that allows for ongoing monitoring and visualisation of risks that have been discovered is easily integrated with the detection outputs. Strong detection capability with high precision and recall while keeping effective frame processing speed is demonstrated via experimental evaluation. The technology offers a workable and scalable solution for smart city infrastructures and intelligent transportation systems. Vigilant RoadShield uses real-time AI-powered monitoring to improve road safety and lower accident-related risks by facilitating automatic hazard identification and quicker response mechanisms. Keywords:Road Hazard Detection; YOLOv8; Real-Time Object Detection; Intelligent Transportation Systems; Deep Learning; Computer Vision; Smart City Surveillance; Road Safety Analytics.