Real-time Triple Seat Detection and Localization of Motorcycle
Keywords:
Convolutional Neural Networks (CNN), Motorcycle, Optical Character Recognition (OCR), Traffic rules, YOLOv8Abstract
Motorcycles, also known as bicycles, are the most widely used and popular means of transportation worldwide. However, following traffic rules can pose significant risks to motorcyclists. Current traffic monitoring systems could be more efficient because they rely on human intervention, which may only sometimes be appropriate. As a result, the rate of road accidents and deaths from such accidents has increased. This project aims to solve this problem by automatically identifying three drivers. The system uses traffic video as input and uses the YOLOv8-based object recognition model to identify motorcycles in traffic scenes. To detect three drivers, the YOLOv8 model identifies individuals and uses a known number of drivers. If the number of motorcycle riders exceeds two, it shows that there is a tricycle ride. A video of the motorcyclist violating the rules is then recorded for the ride.