Intel, Red Cross: Mapping For Disasters With AI

Intel is working with the Red Cross to use artificial intelligence to create highly detailed maps of developing areas that will be ready for first responders to use in the event of a disaster, conflict, or disease epidemic.

Intel is working with the American Red Cross, the Missing Maps project and MILA to bring digital mapping to Uganda.

The Santa Clara, Calif.-based chip maker announced at CES 2020 the collaboration is using artificial intelligence to help identify previously unmapped bridges, roads and cities on satellite images of developing regions such as Uganda.

Intel’s AI Platforms Group Product Manager Matt Beale told CRN, “Unfortunately, while a place like the United States has excellent maps — you can just go to Google Maps — a lot of the developing world has really poor maps or no maps in many cases.”

id
unit-1659132512259
type
Sponsored post

Which means that first responders lack necessary information to make fast decisions regarding relief efforts.

Efforts like the Missing Maps project, put together by humanitarian organizations including the American Red Cross, rely on volunteers to help fill in the blanks of under-mapped areas by gathering geographical data.

Even so, current mapping methods are time-consuming and require a lot of resources.

“Right now the Red Cross rely’s primarily on human volunteers and independently we knew that Intel AI researchers were working on AI and satellite imagery,” Beale told CRN. “So, it seemed like there was this great opportunity to take that research and see how we can augment the Red Cross’ work.”

Fast forward to 2019, the American Red Cross, its Missing Maps project and Intel are using AI to map vulnerable populations in need of disaster planning and emergency response.

“So that’s what we’ve done the past year or so is see how we can work with the Red Cross first in Uganda to identify bridges and essentially help map out vulnerabilities in the transportation infrastructure there using artificial intelligence; not replacing humans, but figuring out how to augment them,” said Beale.

How It Works

Intel told CRN satellite images can be challenging to read, but with the help of an AI model, mappers are able to cover more ground and catch things that may be difficult for the human eye to find.

For example, the model found 70 bridges in southern Uganda that had been missing in previous maps.

“Once we had the trained models we needed to run them at essentially country-wide scale. The advantage of this is to be able to do what would take a human hundreds of hours in seconds or minutes,” Beale said. “So we ran these trained models on 2nd Generation Intel Xeon Scalable processors.”