Wall Crack Image Dataset for Earthquake and Structural Health Analysis
Description
The dataset includes 526 wall crack images which are acquired in diverse actual applications involving residential and commercial building, indoor and outdoor walls and different kinds of wall materials. The scenes were from varied residential and commercial buildings captured by modern smartphone based cameras; the measurements were done manually using digital devices. The images were recorded under different modality conditions including changing in lighting, wall texture ad crack pattern for diversity generation. This data set may help researchers study the origin of wall cracks, earthquake influence on buildings, and practices of structural examination. The dataset can be split by researchers based on their own experimental or methodological conditions. It will be of special interest to students, engineers and researchers in the fields of civil (structural) engineering, structural health monitoring, and disaster evaluation. With offering the variety of wall cracks, we expect such dataset to stimulate the research of earthquakeinduced structural damage and to foster. Keywords: Wall cracks, Structural health monitoring, Civil engineering, Earthquake damage, Image dataset, Structural inspection.
Files
Steps to reproduce
Image Collection: Capture wall crack images from residential and commercial buildings using a standard smartphone camera. Include indoor and outdoor walls with various materials (brick, concrete, plaster, mixed). Measurement Recording: Record relevant measurements and observations manually using digital tools. Image Variability: Ensure diversity by capturing images under different lighting conditions, wall textures, and crack patterns. File Organization: Save each image in JPG/PNG/JPEG/WEBP format with a unique filename Metadata Creation: Prepare a metadata CSV file listing each image along with associated categories such as crack_type, building_type, and material_type. Packaging: Place all images, the metadata CSV, and a README.txt file in a single dataset folder for publication.
Institutions
- Daffodil International University