Dr. Maryam Rahnemoonfar is an Associate Professor of Computer Science and Engineering and Civil and Environmental Engineering at Lehigh University. Her research focuses on developing artificial intelligence and remote sensing technologies for disaster damage assessment and resilience.
The technology has been supported by National Science Foundation research initiatives and developed through collaborations within CERCAT. Engagement with insurance industry partners has helped guide prototype validation, technical development, and the identification of commercialization opportunities, particularly for applications in catastrophe assessment and insurance claims management.
Background
Natural disasters, including hurricanes, floods, wildfires, and severe storms, cause widespread damage to homes, businesses, and critical infrastructure, creating an urgent need for rapid and accurate damage assessments. Timely information about the extent and severity of damage is essential for coordinating emergency response, processing insurance claims, allocating recovery resources, and supporting affected communities.
However, traditional damage assessment methods often rely on manual inspections that are costly, time-consuming, and difficult to scale across large geographic areas. While artificial intelligence offers opportunities to automate these processes, existing models frequently require extensive labeled training data and struggle to adapt to different disaster types, imaging sources, and environmental conditions. These limitations can delay critical decisions when reliable damage information is needed most.
Technology Overview
This technology is an artificial intelligence-driven platform designed to automatically detect, classify, and quantify disaster-related damage using multiple sources of geospatial imagery, including satellite, aerial, and unmanned aerial vehicle (UAV) observations. By combining advanced computer vision, object detection, and image segmentation algorithms, the platform transforms large volumes of remote sensing data into actionable information for disaster response and recovery.
Unlike conventional AI approaches that depend heavily on disaster-specific labeled data, the platform is designed to assess damage across different geographic regions and disaster scenarios without extensive retraining. Initial development and validation have focused on hurricane and flood damage, with demonstrated capabilities in identifying affected structures and infrastructure. Ongoing development aims to expand the platform to wildfires, tornadoes, and earthquakes while integrating Synthetic Aperture Radar (SAR) imagery to improve assessment capabilities under cloud cover, smoke, and adverse weather conditions.
Developed through research and industry collaboration within the Consortium for Enhancing Resilience and Catastrophe Modeling (CERCAT), the technology is being refined to meet the operational needs of insurance providers and disaster response organizations.
Benefits
- Automates the detection, classification, and quantification of disaster-related damage across large geographic areas.
- Accelerates damage assessments by processing satellite, aerial, and UAV imagery, reducing reliance on time-intensive manual inspections.
- Supports assessment of emerging disaster events without requiring extensive event-specific labeled training data.
- Enables insurance providers to identify affected properties, prioritize inspections, and streamline claims processing.
- Provides timely, objective damage information to improve emergency response and recovery resource allocation.
- Offers a scalable platform with the potential to support multiple disaster types and remote sensing technologies.
Applications
- Insurance claims assessment and catastrophe risk management
- Emergency management and disaster response
- Infrastructure damage assessment and utility restoration
- Government disaster recovery and resource allocation
- Humanitarian response and community resilience planning