Intel uses AI and satellite imagery to help the Red Cross map vulnerable locales

Five years ago, the American Red Cross, the British Red Cross, the Humanitarian OpenStreetMap Team, and Doctors Without Borders kicked off the Missing Maps project, a humanitarian endeavor that aims to preemptively map parts of the world vulnerable to conflicts, natural disasters, and disease epidemics. So far, its cartographers have contributed data on towns and cities in countries such as South Sudan, the Democratic Republic of the Congo, and the Central African Republic, which organizations have used to track the movement of displaced populations and to identify the sources of disease outbreaks.

Historically, the bulk of the mapping process has been conducted manually — humanitarian groups identify vulnerable areas and create preliminary maps, after which field staff work with local people to travel to the mapped areas and input the names of key places while checking for errors. Those volunteers are also responsible for tracing swamps, rivers, and other natural landmarks onto satellite imagery, and for submitting the collected info to an online portal so that it can be distributed widely.

In search of a new approach that might help expedite the process, Intel in early 2019 partnered with the Red Cross to provide technology that identifies previously unmapped bridges and roads on satellite images. The organizations say that the AI model “significantly” augments volunteer mappers’ abilities to cover more ground and catch things difficult for the human eye to see.

“[T]here are entire parts of the world that are unmapped, which makes planning and responding to disasters much more difficult,” said cofounder of Missing Maps and CEO of American Red Cross Cascades Region Dale Kunce. “This is why we’re collaborating with Intel to use AI to map vulnerable areas and identify roads, bridges, buildings and cities.”

While mapping Uganda, for instance, Intel’s computer vision model — which runs on 2nd Generation Intel Xeon Scalable processors with DL Boost (an AI workload acceleration suite) and nGraph (an open source AI and runtime/compiler library) — found 70 bridges in the southern part of the country missed by either OpenStreetMap or the Ugandan Bureau of Statistics’ map. “As someone who’s been on the ground with the Red Cross, having access to accurate maps is extremely important in disaster planning and emergency response,” added Kunce.

Intel doesn’t own the complete rights to the mapping corpus it’s helping to create, but the company says that it’s exploring opportunities to open-source parts of it for researchers. Intel also plans to host workshops on the applications of satellite imagery and AI for humanitarian use cases, using the data set and codebase developed for the Missing Maps collaboration.

Intel’s work in AI and mapping comes after Facebook made available Map With AI to the OpenStreetMap community, a tool that automates several of the most time-consuming steps involving in annotating roads, buildings, and bridges. Additionally, the Menlo Park company recently made available TapiD, RapiD, and an AI-powered version of OpenStreetMap’s editing tool iD, in addition to AI-generated road mappings in Afghanistan, Bangladesh, Indonesia, Mexico, Nigeria, Tanzania, and Uganda.

Coinciding with Facebook’s efforts, Microsoft’s Bing Maps team, Microsoft Philanthropies, and the Humanitarian OpenStreetMap Team launched in September an initiative that seeks to bring new machine learning approaches and open building corpora into OpenStreetMap’s auditing tools. As a first step, Bing Maps engineers plan to release country-wide building footprint data sets for Uganda and Tanzania, created with feature-identifying computer vision algorithms.

Source: Read Full Article