Traffic congestion is one of the biggest challenges in modern cities, causing delays, pollution, and inefficient fuel consumption. This project leverages computer vision and deep learning to analyze real-time traffic footage from CCTV cameras. The system uses AI-based object detection models like YOLO (You Only Look Once) or OpenCV to count the number of vehicles on each road. Based on congestion levels, it dynamically adjusts traffic light durations to optimize vehicle flow, reducing unnecessary waiting times. Additionally, the system can detect traffic violations, such as running red lights or illegal lane changes, and automatically notify authorities. This AI-powered system can be integrated with IoT sensors for vehicle counting, and GPS data to prioritize emergency vehicles, such as ambulances and fire trucks, ensuring a smoother and more efficient transportation network.