Road transportation is one of the most important parts of modern infrastructure. Safe and well-maintained roads help people travel efficiently and support economic growth. However, road surfaces often develop damages such as cracks, potholes, and surface wear due to heavy traffic, weather conditions, and aging. Traditional road inspection methods are expensive, time-consuming, and require manual labor. Recent developments in Artificial Intelligence (AI) and Deep Learning (DL) have introduced automated systems for road damage detection. This review paper discusses the use of Vision Transformers (ViT) combined with lightweight Convolutional Neural Networks (CNNs) for detecting road damage from images. The proposed approach improves detection accuracy by capturing both local and global image features. The system achieves approximately 94% accuracy and provides an effective solution for transportation safety, infrastructure monitoring, and smart city development.

Keywords—Artificial Intelligence (AI), Deep Learning, Road Damage Detection, Computer Vision, Convolutional Neural Networks (CNN), Vision Transformer (ViT), Image Processing, Road Infrastructure, Smart Transportation, Pothole Detection.