BD-PaveSurface: A Multi-Weather Pavement Surface Image Dataset for Crack, Pothole, and Good Pavement Condition Assessment in Bangladesh

This dataset contains labeled pavement surface images collected from selected segments of National Highways N5 and N6 in Bangladesh under multi-weather field conditions. The dataset includes three pavement condition categories: Crack, Pothole, and Good (normal/undamaged pavement surface). A total of 21,000 images are provided, comprising 9,000 raw images (3,000 per class) and 12,000 augmented images (4,000 per class), ensuring a balanced class distribution across all categories. Images were captured using a handheld smartphone camera under real-world traffic and natural lighting conditions across rainy season, winter, and summer, and were manually reviewed and labeled for annotation consistency. The dataset was systematically organized and partitioned into training, validation, and testing subsets to support reproducible experiments. This dataset is intended to facilitate research in pavement condition assessment, automated crack and pothole detection and classification, and benchmarking of computer vision and deep learning based pavement inspection methods