Towards Generative Location Awareness for Disaster Response: A Probabilistic Cross-view Geolocalization Approach

ArXi:2512.20056v2 Announce Type: replace As Earth's climate changes, it is impacting disasters and extreme weather events across the planet. Record-breaking heat waves, drenching rainfalls, extreme wildfires, and widespread flooding during hurricanes are all becoming frequent and intense. Rapid and efficient response to disaster events is essential for climate resilience and sustainability. A key challenge in disaster response is to accurately and quickly identify disaster locations to decision-making and resources allocation.