Situational Awareness Detection and Warning for Airport Operations

Abstract

The operation of unmanned aerial vehicles (UAVs) presents several hazardous situations for both ground personnel and vehicles. Collisions between the UAV and other objects can have very negative consequences resulting in potentially high costs as well as injuries or even death of human personnel. When we were approached by Northrop Grumman, they were searching for a way to mitigate these limitations by providing increased situational awareness in the region surrounding their UAV while it was in transit on the ground. After weighing the variety of options available to us we settled on the use of computer vision as our solution because it provided the greatest potential for success with our limited budget. Computer vision works by taking images received from simple digital cameras and processing them to provide the localization information we need. Through our efforts we have been able to design and build a scalable system that can isolate objects and generate both visual and audio warnings when other objects approach too close. The completed system offers an overhead display of the airfield, live video of the scene, scalable network architecture, and clear audio warnings.