ESkyScan Autonomously Filming Aircraft

This talk presents Skyscan, a project from a rapid prototyping group that automatically films aircraft broadcasting their position. The speakers begin with ADS-B, the transponder system used across civil aviation, which sends position, speed and heading in the clear. It can be received with an inexpensive software defined radio and decoder software. Aircraft ownership can be looked up in FAA registry records, with crowdsourced databases as a fallback, though trustees and LLCs often obscure the real owner.

The hardware is modest: a pan-tilt-zoom security camera, a small single-board computer or laptop, and power-over-Ethernet wiring, for roughly $500 with the camera making up most of the cost. The software converts each aircraft's coordinates into azimuth and elevation relative to the camera, then interpolates between sparse and delayed updates so the motion looks smooth. Targets are prioritized by distance or interest, and the system is containerized and published on GitHub. Careful leveling of the camera matters, since small orientation errors affect tracking.

The speakers argued that a long archive of imagery builds a baseline of normal air traffic, which helps analysts spot anomalies such as unexplained drone reports. Labels from FAA records could also be used to train image classifiers for aircraft types. They closed by inviting attendees to build the system and to talk with them afterward.