Struct-BB: Original, Annotated and Segmented Dataset of Exposed Reinforcement in Bridges and Buildings

This Struct-BB Dataset, compiled from real inspection photographs of reinforced concrete bridges and buildings displaying exposed reinforcement. The dataset addresses the scarcity of publicly available segmentation data focused on exposed reinforcement-related pathology, a recurrent damage in concrete structures and an important indicator of structural vulnerability. A total of 64 images were collected during inspections in the Federal District, Brazil, as only these clearly show exposed reinforcement. These images capture real-world variations in illumination, texture, geometry, and pathological features. Each image includes a YOLO segmentation annotation, generated using Trainable Weka Segmentation (Fiji/ImageJ) and manually refined, then converted into polygon-based YOLO files.