AI Driven Helmet-less and Triple-Seat Violation Detection System

This paper presents an automated traffic violation detection system capable of identifying helmetless riders and triple-seat violations using artificial intelligence. The proposed solution integrates the YOLOv8 (You Only Look Once, version 8) real- time object detection framework to analyze traffic video feeds. The system detects motorcycles, riders, and helmets and applies rule-based logic to classify violations. Dataset preparation and annotation are performed using Roboflow, while Python and its supporting libraries form the technological backbone of the implementation. The model processes video frames in real time, flags violations, and generates evidence for enforcement agencies. The system aims to reduce manual monitoring efforts, improve traffic law enforcement, and enhance road safety.