DEVELOPMENT OF A COMPUTER VISION-BASED FRAMEWORK FOR AUTOMATED TRAFFIC SIGN RECOGNITION IN INTELLIGENT TRANSPORTATION SYSTEMS

Abstract

The rapid development of intelligent transportation systems (ITS) has increased the demand for reliable and automated traffic sign recognition to support road safety, traffic management, and autonomous driving applications. However, variations in illumination, weather conditions, viewing angles, sign occlusion, image quality, and complex road environments continue to challenge the accurate recognition of traffic signs. This study aims to develop a robust computer vision-based framework for automated traffic sign recognition within intelligent transportation systems. The proposed framework integrates image preprocessing, traffic sign detection, feature extraction, and deep learning-based classification to automatically identify and categorize traffic signs from road-scene images. A labeled traffic sign dataset is employed for model training, validation, and testing, while performance is evaluated using accuracy, precision, recall, F1-score, mean average precision, and confusion-matrix analysis. The proposed framework is designed to combine reliable recognition performance with computational efficiency so that it can support real-time transportation applications. The expected findings indicate that deep learning-based computer vision can provide high recognition accuracy while maintaining reliable performance across diverse traffic sign categories and challenging visual conditions. The study further highlights the importance of data augmentation, multi-scale feature extraction, and lightweight architectures for deployment on resource-constrained intelligent transportation platforms. The proposed framework can contribute to automated road-scene understanding and support advanced driver-assistance systems and autonomous vehicles. Its practical implications include improved road safety, traffic efficiency, and intelligent mobility.