Industry

AI-generated text

On-site AI analysis of drone imagery speeds disaster response

Researchers at Texas A&M University developed CLARKE, an AI system that analyzes drone imagery offline to produce compact damage reports, reducing the need to transmit large image files over unreliable post-storm networks.

On-site AI analysis of drone imagery speeds disaster response

Researchers at Texas A&M University, with funding from the National Science Foundation’s AI Institute for Societal Decision Making, have developed a system called CLARKE that analyzes drone imagery on-site without requiring an internet connection. The system aims to reduce delays caused by transmitting and interpreting large image files after storms, enabling faster operational decisions and recovery.

How CLARKE works

After a drone lands, CLARKE processes the captured images locally and produces compact, structured reports describing building damage and road conditions. These reports are far smaller than raw image files, so responders can share actionable information over limited or unreliable post-storm networks rather than trying to send huge photographs or video.

Texas A&M researcher Tom Manzini said, “The information exists for emergency managers to make decisions, but it exists in a place where they can’t get to it.” Manzini also noted that CLARKE can assess roughly 2,000 homes in seven minutes.

Deployment and impact

Emergency teams used CLARKE after the 2024 hurricanes Debby and Helene. The system is now incorporated into the standard operating procedures of Florida’s disaster drone team, FL-UAS1. To date, Manzini and his team have trained 125 emergency responders to use CLARKE.

By shortening delays in transferring and interpreting imagery, researchers estimate CLARKE could enable emergency managers to make decisions about a day earlier than they otherwise would. That earlier decision-making allows resources to be directed first to the hardest-hit areas, which in severe crises can save lives.

Related AI efforts

CLARKE is part of a broader movement to apply artificial intelligence in emergency response. For example, Microsoft’s AI for Good Lab analyzed satellite imagery to help aid organizations identify damaged buildings after the March 2025 earthquake in Myanmar.

Why this matters

Drone imagery provides detailed, granular information about damage, but getting usable intelligence to decision-makers quickly can be blocked by damaged or overloaded communications infrastructure. CLARKE’s approach — automated, on-site processing and compact reporting — addresses that bottleneck and can improve the speed and effectiveness of post-disaster response efforts.