ConceptCrack: Autonomous Crack Intelligence for Post-Disaster Structural Assessment
Autonomous AI pipeline fusing YOLO26, SAM 3, and LLM for real-time crack detection and structural risk reporting.
- Crack types
- 6-Class
- Hairline, flexural, shear, diagonal, wide, map cracking
- Segmentation
- SAM 3
- Sub-mm pixel-precise masks
- Reasoning
- LLM
- Risk scoring & remediation
- Standard
- ACI 224R
- Crack width compliance
Overview
ConceptCrack is an end-to-end AI framework for post-disaster structural crack assessment. YOLO26 detects and classifies cracks in real time, SAM 3 delivers pixel-precise segmentation, and an LLM infers cause, assigns risk levels, and generates ACI 224R-compliant remediation reports.
Novel contributions
- 01
Staged AI fusion
YOLO26 proposals feed SAM 3 as spatial prompts; mask statistics feed LLM as structured context.
- 02
Auto-classification
Morphometric rules assign one of 6 concrete crack categories from segmentation masks.
- 03
LLM structural reasoning
Chain-of-thought cause attribution, L1-L5 risk scoring, and prioritized remediation.
- 04
Temporal tracking
Cross-inspection crack geometry comparison supports predictive maintenance planning.
Interested in this research?
For code access, collaboration or questions about ConceptCrack, contact the principal investigator, Dr. Tadesse Wakjira.