Default Dataset for Final Project (Section 002)

Note: if you do not want to use this dataset, you will need to propose your desirable dataset(s) to Dr. Lianzhang Bao for approval before proceeding to develop your project. The goal is to have a dataset (or combo of multiple datatsets) that is complex enough to accommodate a wide range of methods covered in class.

The data: Concrete Cracks Image Dataset

A zip containing concrete cracks image datasets can be download kaggle.

What is it about

Context: Structural Defects Network (SDNET) 2018 is an annotated image dataset for training, validation, and benchmarking of artificial intelligence-based crack detection algorithms for concrete. SDNET2018 contains over 56,000 images of cracked and non-cracked concrete bridge decks, walls, and pavements. The dataset includes cracks as narrow as 0.06 mm and as wide as 25 mm. The dataset also includes images with a variety of obstructions, including shadows, surface roughness, scaling, edges, holes, and background debris. SDNET2018 will be useful for the continued development of concrete crack detection algorithms based on deep learning convolutional neural networks, which are a subject of continued research in the field of structural health monitoring.

Content: 230 images of cracked and non-cracked concrete surfaces (54 bridge decks, 72 walls, 104 pavements) are captured using a 16 MP Nikon digital camera. The bridge decks were located at the Utah State University system, material, and structural health (SMASH) laboratory. The inspected walls belong to Russell/Wanlass Performance Hall building located on USU campus The pavement images were acquired from the roads and sidewalks on USU campus. Each image is segmented into 256 × 256 px sub-images. Each sub-image is labelled as ‘Cracked’ if there was a crack in the sub-image or ‘Non-cracked’ if there was not a crack.

Acknowledgements

Utah State University Campus, Logan, Utah, USA