Open Dataset · The Hashemite University · Jordan

Infrastructure Crack
Atlas for the World

An open-source concrete crack dataset for AI training, structural engineering research, and infrastructure monitoring — collected at The Hashemite University.

41
Crack Records
615+
Total Images
4
Structure Types
CC BY 4.0
License
About the Dataset

Built for the global research community

A systematic, high-quality collection of concrete crack images captured at The Hashemite University campus using a professional protocol designed for AI model training.

10–15 images per crack

Far, medium, close-up from 3 angles, scale reference, 3 lighting conditions, and negative sample.

Standardized field data

Every crack includes type, severity, width, length, surface conditions, and environment metadata.

Open-source forever

CC BY 4.0 license. Free to use, share, and build upon for any research or commercial purpose.

// crack record schema
crack_idCR01 → CR041required
typelongitudinal / transverse / diagonal / hairline
severityminor / moderate / severe
structurecolumn / wall / pavement / beam
width_mmmeasured in millimetersfloat
length_mmmeasured in millimetersfloat
environmentindoor / outdoor
images[]10–15 photos per crackarray
Dataset Gallery

Browse the crack records

Filter by structure type, crack type, or severity. Click any record to view full details and download.

Data Collection Protocol

Standardized for scientific accuracy

Every crack follows a strict photo acquisition protocol ensuring consistency and usability for AI training.

01 / Field Data

Crack Characteristics

Each crack is documented with full physical measurements before photography.

  • Type: longitudinal, transverse, diagonal, hairline
  • Severity: minor, moderate, severe
  • Width and length in mm
  • Surface: dry, wet, or dusty
  • Corrosion and water presence noted
02 / Photography

Camera Settings Logged

All camera parameters recorded for full reproducibility.

  • Camera: DSLR high-resolution
  • Resolution: 4000×3000 px minimum
  • Zoom: 1x / 2x / 5x logged
  • Focus: auto and manual
  • Flash and stability recorded
03 / Photo Types

10–15 Images Per Crack

Systematic multi-view acquisition for robust AI training data.

  • Far view: 2–5 m context shot
  • Medium view: 1 m front
  • Close-up: 0°, +30°, −30°
  • Scale reference with ruler
  • 3 lighting: sun, shadow, flash
  • Negative sample included
File Naming Convention

Consistent naming for every image

[CrackID]_[Type]_[Distance]_[Angle]_[Lighting]_[Number].jpg
Example: CR01_close_30cm_0deg_sun_01.jpg
CrackIDCR01, CR02 … CR041
Typefar / medium / close / scale / negative
Distance30cm / 1m / 3m
Angle0deg / +30deg / -30deg
Lightingsun / shadow / flash
Number01, 02, 03
Community Contribution

Help grow the global dataset

Submit your own crack records from any infrastructure project worldwide. All contributions are reviewed and credited.

Submit a Crack Record

Drop crack photos here or click to browse

JPG, PNG — min 4000×3000 px recommended

Contribution guidelines

Photos must follow the protocol: far, medium, close-up (0°, ±30°), scale, lighting, negative. Minimum 10 images per crack. All images must be original work.

Review process

All submissions are reviewed by the research team at The Hashemite University within 3–5 business days. Accepted contributions are credited in the dataset.

Credit & attribution

Contributors are listed in the dataset README and on this website. Your institution will be credited in every crack record you submit.

Contact

Dr. Yousef Almashakbeh · Civil Engineering
The Hashemite University, Zarqa, Jordan
yousefalmashakbeh@hu.edu.jo

How to Cite

Reference this dataset in your research

If you use this dataset in your work, please cite it using one of the formats below.

APA:
Almashakbeh, Y., Hayarat, I., Momani, D., & Al-Ulaymat, R. (2025). HU Infrastructure Cracks Dataset: An open-source concrete crack image collection from The Hashemite University [Dataset]. https://hu-infrastructure-cracks.org
BibTeX:
@dataset{almashakbeh2025huinfra,
  author = {Almashakbeh, Yousef and Hayarat, Israa and Momani, Dirar and Al-Ulaymat, Ridwan},
  title = {HU Infrastructure Cracks Dataset},
  year = {2025},
  institution = {The Hashemite University, Jordan},
  url = {https://hu-infrastructure-cracks.org}
}
Research Team

The Hashemite University · Civil Engineering

Y

Dr. Yousef Almashakbeh

Department of Civil Engineering

Principal Investigator
I

Israa Hayarat

Department of Civil Engineering

Co-Investigator
D

Dirar Momani

Department of Civil Engineering

Co-Investigator
R

Ridwan Al-Ulaymat

Department of Civil Engineering

Co-Investigator
🏛

The Hashemite University

Zarqa, Jordan — Est. 1995

Host Institution
🌍

Global Contributors

Researchers Worldwide

Community Members