🌏 Dr. Lei Zou's Invited Talk at the Japan Geoscience Union (JpGU) Conference
- Gear Lab at TAMU
- Aug 9
- 2 min read
Dr. Lei Zou was invited to give a talk at the Japan Geoscience Union (JpGU) Conference, presenting on “Geospatial Big Data and AI for Smart Humanitarian Mapping.” In his talk, Dr. Zou addressed one of today’s most pressing global challenges — building communities that can withstand the increasing frequency and severity of natural disasters. He emphasized that humanitarian mapping is a crucial component of resilience-building, enabling the identification of affected areas, assessing infrastructural and social impacts, and understanding humanitarian needs during disaster events.
Drawing on advances in geospatial big data—including satellite imagery, street-view images, social media, and crowdsourced platforms—combined with cutting-edge artificial intelligence (AI), Dr. Zou presented a comprehensive framework for Smart Humanitarian Mapping. This interdisciplinary approach enables rapid, accurate, and detailed assessments to support disaster response and resilience. Through three case studies, he demonstrated the practical application of these methods:
Mapping rescue requests during Hurricane Irma using social media and fine-tuned large language models (VictimFinder and TopoBERT).
Mapping power outages during Winter Storm Uri using nighttime light remote sensing with statistical corrections to account for viewing angle and snow reflection effects.
Assessing street-level damages after Hurricane Milton using bi-temporal street-view imagery and dual-channel vision models for hyperlocal disaster loss estimation.
The results show the remarkable potential of integrating diverse data sources and advanced AI algorithms to provide actionable insights during crises, from locating victims in need to identifying power outages and infrastructure damages. By addressing technical and operational challenges in applying these technologies, this research paves the way toward more resilient communities worldwide.
