Real-Time Smart Traffic Violation Detection and Analytics Using a Multi-Model YOLO Framework

Road accidents often stem from breaking traffic rules, especially when drivers ignore red lights at junctions. Inefficient oversight happens with human monitors, which struggle to keep up in today’s busy cities. Instead of relying on people, a method driven by deep learning and visual computing analyzes video from dashcams, security cameras, or still images to catch these red-light breaches instantly. Different versions of the YOLO framework handle distinct tasks - spotting signals, finding cars, tracing their movement, detecting pedestrian crossings, and locating license plates. Rather than fixed lines, the system calculates where vehicles should stop by examining crosswalk positions over time, refining its estimate through sequential frame analysis for better accuracy. From captured footage, the system pulls out license plate details using optical character recognition. After detection, it stores organized entries for each incident. Visualization happens instantly through a live-updating interface built with Streamlit. Analytics appear alongside incoming data, updated without delays.

Keywords: Traffic Violation Detection, YOLO, Deep Learning, License Plate Recognition, Computer Vision, Smart City, Intelligent Transportation.

Deepak M, Gokul S, Giriprasanth S, Vaijayanthi M (2026). Real-Time Smart Traffic Violation Detection and Analytics Using a Multi-Model YOLO Framework. International Journal of Innovative Research in Multidisciplinary Education & Technology (IJIRMET), 11(3).