AI Based Traffic Signal Classification System

The accurate and real-time classification of traffic signals is a safety-critical task for both human drivers and emerging autonomous systems. Misinterpretation of a traffic light can lead to catastrophic accidents. This paper presents an Artificial Intelligence (AI) based system designed for the robust classification of traffic signals from image data. The core of the system is a deep Convolutional Neural Network (CNN), which is trained to automatically learn the distinguishing features of traffic light states (Red, Yellow, Green) under various real-world conditions. We employ a transfer learning approach, fine-tuning a pre-trained CNN architecture on a specialized traffic signal dataset. This methodology allows the system to achieve high classification accuracy while being computationally efficient. The system is evaluated on its ability to perform reliably across challenging scenarios, including varying weather, time of day, and partial occlusions. The results demonstrate the potential of this AI- based approach to serve as a vital perception component for Advanced Driver-Assistance Systems (ADAS) and fully autonomous vehicles.