New paper on AI-assisted visual inspection
A methodological contribution on integrating YOLO-based detection with engineering-constrained prompting for defect description.
See publications →Human-driven AI for Resilient Infrastructure
HAIRI is an academic laboratory at the University of Basilicata focused on integrating artificial intelligence, structural engineering, and risk-based approaches for safer and more resilient civil infrastructure.
AI-based Visual Inspection
Computer vision for bridges and infrastructure: guardrails, bearings, and surface defects.
Engineering-Constrained ML
Machine learning aligned with engineering principles, interpretability, and safety.
Risk & Resilience
Risk-based prioritisation of interventions for road networks and critical assets.
Mission pillars
What we do
Engineering-centred AI workflows with clear subdivisions between people, data and decisions.
Human-in-the-loop AI
We design AI workflows where the engineer remains central: validation, error control, and decision-making are never delegated blindly to algorithms.
From images to decisions
From UAV, street-level, and on-site images to quantitative indicators that support maintenance planning and safety verification.
Open science
Whenever possible, we release datasets, source code and reproducible workflows to support the scientific and professional community.
Updates
Latest highlights
Research outputs, applied projects and events where the HAIRI Lab community is involved.
ARGUS, APPOGGI and VIADUCT
A suite of AI-based procedures for assessing guardrails, bearings and surface defects on bridges and viaducts.
Explore projects →Seminar series on AI & Infrastructure
Regular seminars for engineers, students and authorities on practical applications of AI to infrastructure management.
Upcoming events →
Need a deep dive into our infrastructure AI toolchain?
Connect with the team