Proceedings of the 82nd Military Operations Research Society (MORS) Symposium, Infrastructure Analysis, Protection and Recovery (IAP&R) working group, Alexandria, VA

(June 2014)

Large-scale natural disasters have the potential to create loss of life and property on a large scale. Responding to these incidents is a complex process involving many resources over extended time periods. Determining the number of people and amount of critical infrastructure affected immediately following an incident is critical to gain situational awareness and coordinate an effective response plan. Initial reports of disaster damage are often understated, and predictive computer models often suffer from systematic bias and can vary significantly based on subtle changes in weather patterns. Response planners do not typically receive reasonably accurate assessments of the situation on the ground until airborne impact assessment teams move into isolated or cut-off areas. In this paper, we describe a systematic approach that we are developing to provide rapid evaluation of post incident damage to support faster and more accurate disaster assessments and help direct valuable assets to the most impacted areas, potentially saving lives. Our concept includes multi-objective unmanned aerial vehicle routing, event-specific damage detection, real-time disaster imagery registration, and a collaborative user interface to help enhance situational awareness. We combine advanced object detection algorithms to evaluate georegistered imagery collected in real-time from aerial reconnaissance routes. Applications include disaster assessment and relief, along with supporting local and international humanitarian relief missions.