Automating Disaster Relief

The GEOINT Community endeavors to automate damage assessments following natural disasters

Image

Natural disasters, often unpredictable and devastating, are unavoidable.  But the GEOINT Community is developing a more effective way to respond to them.

On Sept. 1, Hurricane Dorian made landfall on the Abaco Islands in the Bahamas. It hovered above the country for over 48 hours, devastating parts of the archipelago. Two days after the storm, the National Geospatial-Intelligence Agency (NGA) uploaded general mapping data, and five days later, 80 unique products to aid disaster relief efforts.

Satellite imagery can be produced quickly following a natural disaster, but that data still needs to be evaluated by human analysts to determine the extent of the aftermath. Automating that process could be the key to delivering damage assessments to first-responders even faster.

The Pentagon’s Joint Artificial Intelligence Center (JAIC), Software Engineering Institute researchers, the Defense Innovation Unit (DIU), and several other government organizations created the xView2 Challenge, a follow up to xView1, as a $150,000 prize competition to develop algorithms capable of identifying and labeling damage assessments from satellite imagery.

“DIU’s goal in hosting this challenge is to enlist the global community of machine learning experts to tackle a critically hard problem: detecting key objects in overhead imagery in context and assessing damage in a disaster situation,” Mike Kaul, DIU’s AI portfolio director, said in a statement announcing the challenge in August.

xView2 participants use a publicly available dataset, released Sept. 19, of satellite imagery that includes 45,361 square kilometers of pre-disaster and post-disaster imagery of six types of disaster—such as wildfires, landslides, earthquakes, and flood damage.

The DIU reported that approximately 3,000 people have signed up to participate. Their proposed algorithms must quickly and accurately locate and assess the extent of the damage. Submissions are due Nov. 22. Winners are invited to present their work at the December NeurIPS 2019 conference and will be considered eligible for follow-up work with the Defense Department.

“The JAIC is helping to fund the challenge and helped develop the ideas and the concept for the computer vision challenge,” Lt. Cmdr. Arlo Abrahamson, a JAIC spokesman, told C4ISRNET. “Our Humanitarian Assistance and Disaster Relief [HA/DR] team will work closely with the DIU to assess the ideas and recommendations generated from this challenge.”

The winning ideas will contribute to the advancement of the JAIC HA/DR national mission initiative to support and enable the DoD and our agency partners with AI solutions in this important field of work. Partner organizations include NASA’s Earth Science Disasters Program, the Federal Emergency Management Agency’s Region 9, California Governor’s Office of Emergency Services, Cal Fire, the California National Guard, DoD’s Joint Artificial Intelligence Center, Carnegie Mellon’s Software Engineering Institute, the United States Geological Service, NGA, and the National Security Innovation Network.

Using Cutting-Edge Technology for Precision Agriculture

Technological advances in remote sensing for both satellites and UAVs have made them crucial tools in agricultural markets to boost productivity and profitability

The Principles of AI

IC provides new guidance on how agency personnel should develop and use AI and machine learning as part of their intelligence-gathering responsibilities

Monitoring the Earth

The combination of EO satellite data with open-source tools for managing and leveraging that data are leading to more and more countries using EO satellite data support policy-making