Abstract:
Traffic management is a major challenge in densely populated countries like Bangladesh, where traffic violations are common and hard to enforce. In this project, we propose an Development of Intelligent Traffic Management System (DITMS) that uses computer vision techniques to automatically detect and record traffic infractions. The DITMS consists of four modules: (1) License Plate Detection using YOLOv8, a deep learning model that can identify and extract license plate numbers from images; (2) Speed Measurement with Kalman Filter, a mathematical method that can estimate the speed of vehicles from consecutive frames; (3) Vehicle Type Detection utilizing a Convolutional Neural Network (CNN), a machine learning model that can classify vehicles into different categories based on their shape and size; and (4) Vehicle Counting through a combination of object detection, tracking, and segmentation algorithms, which can count the number of vehicles passing through a given area. The DITMS can be deployed on roadside cameras or drones to monitor traffic flow and capture evidence of violations such as speeding, overloading, or illegal parking. The DITMS aims to digitize traffic infractions in Bangladesh and contribute to the vision of SMART Bangladesh 2041, a national initiative that seeks to transform the country into a digital and developed nation
