While so much of the news coverage about drones lately has focused on their negatives, like interfering with air traffic and the use of armed drones, they have also quietly been finding far more positive application in the humanitarian sector.
The United Nations launched its first unarmed surveillance drones in December 2013, flying them over the Democratic Republic of Congo and then Rwanda. The UN drones are used not only for actual surveillance and monitoring tasks, but also to hover at low altitude in full visibility of hostile fighters, as a deterrent to remind them they are being actively observed.
Yet, it is the experimental applications of drones for humanitarian disaster relief that are pushing the boundaries of how aerial imagery technology can reshape our ability to respond to disasters rapidly. Early applications using large volunteer teams to identify wildlife in satellite imagery and train computer algorithms to conduct basic wildlife censuses have evolved into large-scale disaster triage efforts using drone imagery. Under one pilot application, live drone imagery would be streamed to a remote team of volunteers who would click on the video feed to identify damaged locations. Areas receiving large numbers of clicks would be highlighted on the screen for drone operators to investigate further, offering realtime feedback. During one test over 100 volunteers collectively provided 49,706 identifications that were 87% accurate.
Given the enormous volume of imagery that can result from just a single drone flight, significant advances are occurring in the algorithmic assessment and identification of aerial imagery. Earlier this year drone imagery was used to construct an interactive 3D model of a refugee camp. This work has continued to develop, with a video released last week showcasing 3D reconstruction of key earthquake-damaged areas of Nepal.
All of these approaches have long histories in the classified imagery intelligence community, in fields with acronyms like GEOINT, IMINT, and MASINT. What makes their emerging humanitarian deployment so exciting is that as the civilian sector has gained access to the same kinds of high-resolution taskable satellite and realtime drone imagery, the landscape of potential applications and technological solutions has exploded. In a world where you can browse your neighbor’s backyard or walk around Times Square without ever leaving your desktop, the ability to zoom into a hilltop in conflict-stricken Eastern Ukraine is almost taken for granted.
With civilian drone deployments becoming increasingly routine in disaster areas, the rise in the availability of aerial imagery is prompting an influx of fresh ideas and technological approaches. Especially exciting are the potential applications of autonomous flight and coordination technologies. A recent MIT prototype drone demonstrated the ability to autonomously navigate through a lightly forested area at over 30 miles per hour, at a cost of just $1,700 to build and weighing just one pound.
The Resilience Post
Why video games are obsessed with the apocalypse! For decades, video games have had a fascination with the end of the world. Why is it that we find it so […]
If you think an eclipse means doomsday, you are not the first! There’s a long and fascinating history of people believing that eclipses are a sign of the apocalypse. (Hopefully […]
A year ago Canada dodged a terrorist bullet when the almost 25-year old Muslim convert Aaron Driver climbed into a cab outside his sister’s home in Strathroy, a small town […]
Using autonomous flight and coordination capability, the drones would all take to the skies, fanning out in grid formation to image the entire affected area. A subset of drones would use base maps to prioritize critical infrastructure like roadways and hospitals, flying the length of all major roads, rail lines, and other transport corridors and generating updated transit network data that can be uploaded in realtime to logistics planning software to coordinate relief delivery. Using neural network image recognition and online comparison with previous base imagery, hotspot analysis could rapidly identify all of the areas with potential damage.