Two-Wheeler Classification using Deep Learning (YOLOv5)

The creation of an Adaptive Traffic Management System specifically designed for Hyderabad's traffic control framework is aided by this study. The research focuses on the categorization of two-wheeler vehicles in Hyderabad's busy metropolitan environment as a crucial component of this larger project. This analysis is based on real-time data kindly given by the Hyderabad Traffic Department. With the use of non-maximum suppression and You Only Look Once v5 (YOLOv5) for object recognition, this work offers a reliable method for classifying and identifying two-wheelers in real-time. A comparison with Faster Region-based Convolution Neural Network (R-CNN) demonstrates how well the model performs, with improved classification accuracy and significant inference time improvements. This study establishes the foundation for incorporating cutting-edge technologies, demonstrating the practical viability of YOLOv5 for real-time object detection in the optimization of traffic management strategies.