This dataset contains 2700 structural images containing damage: crack and spalling. It is labeled by Cal Unmanned Research Lab from University of California, Berkeley. The source of data base includes public dataset, SDNET and PEER Hub Image Set, and self-collected data.
This dataset labeled crack and spalling with 2D bounding box. labelImg is a good tool to visialize or edit images and labels.
Xin Peng, Gaofeng Su, ZhiQiang Chen, Raja Sengupta. Structural Damage Detection, Localization, and Quantification via UAV-based 3D Imaging, 13th International Workshop on Structural Health Monitoring, USA
This work is supported by Pacific Earthquake Engineering Research (PEER) center project #18081. The inspected concrete columns are supplied by PI Dr. Mohamed Moustafa with the American Concrete Intuition Foundation Concrete Research Council 2019 research grant. The data is collected with the help of Dr. Luna Ngeljaratan and Dr.Mahmoud Aboukifa.