This dataset contains labeled pavement surface images collected from an urban segment of National Highway N6 in Pabna District, Bangladesh. The dataset includes three pavement condition categories: alligator cracking, edge breaking, and normal pavement. A total of 24,000 images are provided, comprising 12,000 raw images and 12,000 augmented images, with balanced class distribution. Images were captured under real-world traffic and lighting conditions using a handheld smartphone camera and systematically organized into training, validation, and testing subsets. The dataset is intended to support research in pavement condition assessment, automated distress detection, and benchmarking of data-driven image analysis methods