Document Type
Article
Journal/Book Title/Conference
13th National Conference on Earthquake Engineering
Publisher
Earthquake Engineering Research Institute
Location
Portland, OR
Publication Date
3-3-2026
Journal Article Version
Version of Record
First Page
1
Last Page
5
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Abstract
Post-earthquake damage assessment is crucial to evaluate the safety and serviceability of bridges after an earthquake. Conventional screening methods rely on manual inspection, which is costly, time-consuming, and poses safety risks to inspectors. To address these challenges, this study presents a context-aware vision transformer framework for integrating structural element details into multi-defect identification through a late-fusion approach. The model is trained on a post-earthquake image dataset collected from three major earthquakes in Italy, consisting of 356 images. Five structural elements of reinforced concrete bridges are selected, where concrete spalling and reinforcement corrosion were chosen as target defects. The framework achieved F1-scores exceeding 80% and 75% for bridge structural elements and defects, respectively. In addition, through its late-fusion process, the model localized the defects to specific elements of bridges, providing accurate inference on unseen test images and suggesting the framework's generalizability.
Recommended Citation
Bakhtiari P, Ratna Pandey S, Zaker Esteghamati M. A Context-Aware Computer Vision-based Approach for PostEarthquake Damage Assessment of Bridge Inventories. Proceedings of the 13th National Conference in Earthquake Engineering, Earthquake Engineering Research Institute, Portland, OR. 2026.
Comments
This article will be rereleased by EERI in July 2026 as part of the full conference proceedings. Published with permission.