Ecuador Pavement Damage Dataset (EPDD)

The Ecuador Pavement Damage Dataset (EPDD) is the first publicly available annotated image dataset for pavement distress detection collected across multiple cities and regions of Ecuador. It comprises 1,295 images (after quality filtering) captured on urban roads, covering three natural regions: Sierra, Costa, and Amazonía, across seven cities: Ambato, Quito, Loja, Santo Domingo, Manta, Puyo, and Zamora. The dataset contains six distress classes: Pothole, Block Crack, Longitudinal Crack, Transverse Crack, Patching, and Alligator Crack. Annotations are in YOLO format. Intended for pavement distress detection research and cross-domain transfer learning experiments.