Competitors vying for the $150,000 prize have to develop computer vision algorithms to automate the process of assessing building damage after natural disasters.
WASHINGTON — The Defense Innovation Unit announced Sept. 25 it has posted the satellite images for its second prize challenge known as xVIEW2, which focuses on using artificial intelligence to assess damage from natural disasters.
Competitors vying for $150,000 in prizes have to develop computer vision algorithms to automate the process of assessing building damage after natural disasters such as earthquakes, tsunamis, floods, volcanic eruptions, wildfires and wind.
DIU, a Defense Department technology outreach office based in Silicon Valley, has posted high-resolution images of 550,230 buildings from 10 different countries obtained from Maxar’s DigitalGlobe Open Data Program. The overhead satellite images show buildings before and after a natural disaster and have annotated polygons and damage scores for each building. DIU said this is one of the “largest and highest quality public datasets of annotated high-resolution satellite imagery.”
Mike Kaul, DIU Artificial Intelligence portfolio director, said in a statement that his organization is looking 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.”
DIU said more than 3,000 people have signed up to compete in the challenge. Contestants will be judged on how quickly and accurately their algorithms can locate and assess the extent of the damage. The goal is to help automate the analysis of imagery and help government agencies expedite relief to disaster areas. When large areas are affected by natural disasters, the huge numbers of pixels representing those areas make it time consuming for analysts to search and evaluate specific buildings.
The competitors’ algorithms will be tested against a new dataset, called xBD, created by experts from academia and industry. According to DIU, the xBD is currently the largest and most diverse annotated building damage dataset used to generate and test models for automating building damage assessments.
The competition ends November 22. The winners will be invited to present their work at the December NeurIPS 2019 Workshop on artificial intelligence for humanitarian assistance and disaster relief. Winners also will be considered eligible to be awarded follow-on work with the Defense Department.
Partner organizations in this challenge 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, the National Geospatial Intelligence Agency and the National Security Innovation Network.
This year’s competition builds upon the xVIEW1 challenge held in 2018, which sought out computer vision algorithms to locate and identify distinct objects on the ground that are useful to first responders.