National Technical University of Athens (NTUA)
Reshaping urban resilience: trustworthy AI-driven frameworks for strengthening public health, critical infrastructure and crisis management from climate change effects
Abstract
dc:descriptionClimate Change (CC) hazards pose escalating threats to urban environments, affecting infrastructure, public health and overall environmental sustainability. To effectively mitigate their effects and with the ultimate goal of enhancing urban resilience, Artificial Intelligence (AI) and Remote Sensing (RS) are widely employed in the information communication technology infrastructures by urban planners to strengthen disaster management procedures. In response to the continuously growing demands to learn, adapt and handle complex, unstructured problems in urban environments, this PhD dissertation couples RS, AI, and eXplainable AI (XAI) to contribute to four main pillars being critical for urban resilience. In the first pillar, public health and urban dynamics monitoring, LIME and SHAP XAI techniques are used to intelligently link environmental parameters with public health data, including demographic and socioeconomic factors, enabling the understanding of complex spatio-temporal transmissions in epidemiological models, emphasizing on COVID-19 morbidity and mortality rates. In the second pillar, urban infrastructure monitoring for damage assessment, Deep Learning (DL) architectures based on U-Net and its variations are explored for rapid and cost-effective assessment of urban infrastructure by automatically detecting and localizing buildings. In the third pillar, focused on land use and land cover (LULC) monitoring, an XAI-enabled framework is introduced that employs several deep learning classifiers alongside SHAP to identify optimal spectral band combinations for improved LULC classification, including pixel-level explanations. In the fourth pillar, Environmental Data Management for Crisis Response, the development of a context-aware adaptive Data Cube (DC) structure is presented. It is an analysis-ready structure of diverse sensing sources harmonically aligned under a unique coordinate system and is adaptable to the varying requirements of different environmental crisis scenarios. This is achieved by incorporating data acquisition, processing, and AI-driven decision-making layers along with Earth Observation (EO) data, ensuring that environmental monitoring and crisis management are both precise and responsive and enhancing the ability to address the challenges posed by CC hazards and natural disasters. Together, these pillars directly address the immediate needs of disaster response and urban health monitoring, setting the foundation for long-term sustainable development and planning in urban environments with utmost goal of reshaping urban resilience.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Temenos, Anastasios
- Τέμενος, Αναστάσιος
Subjects
dc:subject × 30- Τεχνητή νοημοσύνη
- Τηλεπισκόπηση
- Επεξηγήσιμη τεχνητή νοημοσύνη
- Μηχανική μάθηση
- Βαθιά μάθηση
- Διαχείριση καταστροφών
- Αστικός σχεδιασμός
- Χρήσεις Κάλυψης Γης
- Παρατήρηση γης
- Κύβοι Δεδομένων
- Artificial intelligence
- Remote sensing
- Explainable AI
- Machine learning
- Deep learning
- Disaster management
- Urban planning
- Land Use and Land Cover
- Earth observation
- Data Cubes
- Γεωεπιστήμες και Επιστήμες Περιβάλλοντος
- Φυσικές Επιστήμες
- Παγκόσμια και πλανητική αλλαγή, κλιματική αλλαγή
- Περιβαλλοντική Μηχανική
- Επιστήμες Μηχανικού και Τεχνολογία
- Earth and Related Environmental Sciences
- Natural Sciences
- Global and planetary change, climatic change
- Environmental Engineering
- Engineering and Technology
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/60053