Ground-Level Building Damage Images from the 2023 Türkiye Earthquake Sequence

Türkiye earthquake reconnaissance yielded 675 annotated ground-level images of building damage. Domain experts from the Geotechnical Extreme Event Reconnaissance (GEER) Association and the Earthquake Engineering Research Institute (EERI) collected the images during post-event field missions. Each image is classified into one of three damage states: no damage, mild damage, or severe damage.

Use Cases

• Training computer vision models for automated damage classification based on the three annotated damage states.
• Benchmarking machine learning algorithms for rapid post-earthquake building assessment using expert-collected field imagery.
• Studying patterns of structural failure and damage severity in a major seismic event based on ground-level photographic evidence.

Strengths

• Contains 675 expert-collected field images, providing real-world visual data.
• Annotations follow a standardized classification framework (EID PRJ-5748), ensuring consistency with prior datasets.
• Data was collected by domain experts from recognized organizations (GEER and EERI).

Limitations

• The specific distribution of images across the three damage categories is unknown.
• Column-level documentation and metadata for individual images are not provided in the input.
• The geographic precision and exact collection dates within the 2023 event are not specified.

Provenance

Source

Geotechnical Extreme Event Reconnaissance (GEER) Association and Earthquake Engineering Research Institute (EERI).

Collection Method

Collected in the field by domain experts during post-earthquake reconnaissance missions.

Time Range

2023

Freshness

Last updated 2026-06-04 08:10:07.

Geography

Türkiye (areas affected by the 2023 earthquake sequence).