When airfields are damaged during conflict, rapidly restoring them to operational status is critical. Our goal in introducing damage‑detection algorithms into our software was to enhance automation for identifying runway damage, enabling rapid airfield repair with minimal human intervention. We continue to refine and expand software solutions that process live drone video feeds to identify damage and unexploded ordnance (UXO), automatically generating detailed assessments with precise geospatial coordinates that feed directly into repair planning tools.
Traditionally, airfield damage assessment required personnel to manually review reconnaissance imagery; a time‑consuming process that delays essential repair operations. The technical challenge lies in receiving high‑resolution video streams from drones flying over damaged runways, translating that raw video into a format optimized for machine‑learning processing, and presenting detected damage to users for validation and decision‑making, all while maintaining accuracy and reliability in real-world field conditions.
Our new solution uses AI‑powered, containerized software to automatically process drone video in real time. As a drone streams video from a damaged runway, machine‑learning detection models identify damage features with associated confidence scores, extract geospatial coordinates from embedded metadata, and push results directly into our GeoExPT planning software. This capability reduces assessment time, minimizes personnel exposure to hazardous areas, and enables repair teams to make faster, better‑informed decisions – ultimately accelerating restoration of operational capability to the airfield.
