Analysing over 2,000 buildings affected by the 2020 Beirut explosion and 93 structures impacted by the 2023 Kakhovka Dam breach, this talk presents data-driven methodologies for assessing heritage building damage from anthropogenic disasters. In both cases, our machine learning models achieved over 90% accuracy in predicting damage levels, illustrating the effectiveness across different disaster types and scales of available data. Key building attributes influencing vulnerability were identified, including urban setting, occupancy, age, structural characteristics, and topography. Some factors like soft stories and structural modifications showed limited impact on explosion damage severity in Beirut, while factors such as foundation type significantly influenced damage in the case of the dam breach. These findings illustrate the importance of context-specific vulnerability assessments. To streamline data collection, we developed a workflow that integrates 2D image segmentation with 3D photogrammetric point clouds; this approach enables rapid detection and quantification of multiple distinct damage types, facilitating further efficient data-driven studies. By combining computer vision, machine learning, and 3D modelling techniques, this work provides tools for enhancing heritage building resilience assessment across different anthropogenic disasters. The results from this work can inform policy decisions, guide targeted retrofitting strategies, and improve disaster preparedness and response planning.
The Terrorism Risk Assessment, Modelling and Mitigation Seminar Series (TRAMMSS) is a virtual seminar series focused on technical topics related to terrorism risk assessment, and modelling, including blast modelling and response; IEDs; vehicles as weapons; CBRN; big data for risk assessment, security and screening; and associated mitigation measures.
Speakers
Joe Kallas is an architect and cultural heritage specialist with experience in disaster response and reconnaissance missions. As an expert member of CIPA Heritage Documentation, he utilises digital technology to support recovery efforts in post-disaster areas. Also serving as a UNESCO international expert, Joe contributes to heritage preservation in disaster-affected areas such as Beirut, Ukraine, and Syria. He is currently a PhD candidate at Penn State University, where he focuses on integrating artificial intelligence and machine learning with 3D structural assessments to improve disaster reconnaissance and enhance building resilience.
Dr Rebecca Napolitano, Assistant Professor of Architectural Engineering at Penn State University, leads the BEAM lab, which specialises in reconciling transdisciplinary data about historic buildings into actionable information. With a PhD in Civil Engineering from Princeton University, Dr Napolitano brings an interdisciplinary perspective to heritage preservation and disaster resilience research. The BEAM lab researches automated damage detection, quantitative data extraction from structural assessments, and the synthesis of heterogeneous information sources.
Who should attend
This seminar is open to guests from outside 星空体育官网, who may work in academia, research, or industry. Due to the potentially sensitive nature of this seminar series, guests should be able to show that they are affiliated with an appropriate bona fide organisation.
Cost
The event is free of charge, but participants must register for the TRAMMSS mailing list in advance.How to register
To attend this seminar, you must register for the TRAMMSS mailing list via the .
Further information on the TRAMMSS community can be found on the main website.
