Overview
Graph theory has a long and varied history of being applied in various fields of knowledge. In this particular project, a representative graph is used to model the entire pattern of cracks on a concrete surface. The plan is to use computer vision techniques to process images of concrete surfaces with cracks. The cracks will be identified and extracted from the images and represented as a graph, which is a mathematical model suitable for computers.
The mathematical principles of graph theory are applied to quantify the crack patterns. For this, the graph theory’s extensive literature is used to describe the key features of each graph with only a few numerical values. These numerical values are called graph features that will be used to quantify the crack patterns. In the subsequent stage of this data processing pipeline, machine learning algorithms will be developed based on the graph features to perform predictive analytics. The project aims to predict structural damage by utilizing information derived from the structures’ appearance.
GitHub Repository
Published Peer-reviewed Papers
- Pedram Bazrafshan, Thinh On, Sina Basereh, Pinar Okumus, and Arvin Ebrahimkhanlou, “A graph-based method for quantifying crack patterns on reinforced concrete shear walls,“ Computer-Aided Civil and Infrastructure Engineering 39, No. 4 (Feb. 2024), 498-517
https://doi.org/10.1111/mice.13009- Featured as the cover of the journal‘s special issue on computational concrete engineering in February.
Awards
- O. H. Ammann Research Fellowship in Structural Engineering by the American Society of Civil Engineers (ASCE), 2024
Media Coverage
- Featured as the Cover of the journal‘s special issue on computational concrete engineering in February.
- Ammann Fellowship Awarded to Pedram Bazrafshan
- Reading Between the Cracks
- Homeland Security News Wire
- https://bioengineer.org/reading-between-the-cracks-artificial-intelligence-can-identify-patterns-in-surface-cracking-to-assess-damage-in-reinforced-concrete-structures/
- https://www.bug.hr/istrazivanja/laboratorijski-uzgojeni-tumori-mogu-predvidjeti-ishode-lijecenja-38126
- https://obrasurbanas.es/la-ia-podria-identificar-el-colapso-de-una-estructura-de-hormigon/
- https://newsazi.com/artificial-intelligence-can-identify-patterns-in-surface-cracking-to-assess-damage-in-reinforced-concrete-structures/#google_vignette
- https://techxplore.com/news/2023-06-artificial-intelligence-patterns-surface-concrete.html#google_vignette
- https://www.miragenews.com/artificial-intelligence-identifies-patterns-in-1018530/#google_vignette
- https://www.eurekalert.org/news-releases/991129
Conference Papers
- Pedram Bazrafshan and Arvin Ebrahimkhanlou “Detection of cracking mechanism transition on reinforced concrete shear walls using graph theory”, Proc. SPIE 12950, Nondestructive Characterization and Monitoring of Advanced Materials, Aerospace, Civil Infrastructure, and Transportation XVIII, 129500I (9 May 2024)
https://doi.org/10.1117/12.3011092 - Bazrafshan, P., & Ebrahimkhanlou, A. (2023). A robotic-based framework for quantifying surface cracks of concrete shear walls. STRUCTURAL HEALTH MONITORING 2023
https://www.dpi-proceedings.com/index.php/shm2023/article/view/36868 - Pedram Bazrafshan and Arvin Ebrahimkhanlou “A virtual-reality framework for graph-based damage evaluation of reinforced concrete structures”, Proc. SPIE 12487, Nondestructive Characterization and Monitoring of Advanced Materials, Aerospace, Civil Infrastructure, and Transportation XVII, 1248707 (18 April 2023)
https://doi.org/10.1117/12.2657736 - Pedram Bazrafshan, Thinh On, and Arvin Ebrahimkhanlou “A computer vision-based crack quantification of reinforced concrete shells using graph theory measures”, Proc. SPIE 12046, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2022, 120460J (18 April 2022)
https://doi.org/10.1117/12.2612359
Crack-to-graph Converted Samples
Sketch of The Crack Patterns
Representative Graphs
