JACKSONVILLE, Fla.—Using small third-party drones for high-resolution imaging and its own artificial intelligence-based software algorithms, Boeing is expanding an effort to speed the inspection of military aircraft exteriors and increase readiness while reducing costs and enhancing safety, a company official said.
Boeing has developed what it calls “automated damage detection software” that quickly analyzes video taken by a small drone hovering above and around an aircraft to automatically detect everything from chipped paint and exposed screws and rivets to missing seals and exposed composite fibers, Scott Belanger, who works contested logistics solutions for Boeing’s Global Services segment, told reporters.
The current manual inspections of large aircraft like a KC-46 tanker and C-5 transport, which have tails that stand 51-feet and 72-feet high respectively, are difficult to conduct and dangerous and “not accurate even when you try your best with the human eyeball,” Belanger said during a June 27 briefing in a hangar where the company does maintenance, repair, and overhaul work on military aircraft.
Officials from Boeing and its partner Skydio, a developer and manufacturer of small quadcopter drones, demonstrated the autonomous aircraft general visual inspection capability on a moveable step platform standing about 10-feet high representing a small aircraft. For the demonstration, a Skydio official remotely flew his company’s X10D drone out a few dozen yards and marked off a zone around platform, then touched the display screen on his controller to enable the small unmanned aircraft to autonomously fly above and around the platform at until it finished imaging with an onboard 64-megapixel Teledyne FLIR camera.
The corrosion problem facing legacy Defense Department aircraft is tens of millions of dollars monthly and Boeing wants to help cut into those costs, said Belanger, who served in the Air Force and was a maintenance commander. A typical “home station” inspection by the Air Force of a large aircraft like a KC-46 or C-17 transport takes about six hours and Boeing and its partners are halving that time with the combination of the small drones and anomaly detection software, he said.
In 2023, Boeing’s autonomous aircraft inspection program collected over 4,000 images of different aircraft and achieved “a 93 percent true positive detection rate,” which means a human technician validated that the software correctly identified a corrosion issue in those instances, Belanger said. Depending on the experience of the personnel, an inspection team will miss about 50 percent of the damage, he said.
Moreover, the drones provide inspection angles that inspection teams cannot obtain manually and allow a customer to build a consistent and accurate digital record of the aircraft that they are not currently getting, Belanger said. And every time Boeing runs its software against the data, the company’s algorithm learns and gets better, he added.
Boeing initially partnered with Near Earth Autonomy, which provides small drones equipped with LiDAR sensors, for the autonomous aircraft inspection effort. Recently, Boeing expanded its partners to include Skydio, which has shipped more than 40,000 drones worldwide, including 2,000 to Ukraine.
Aircraft that have been examined with the drones and AI software include Boeing’s KC-46, KC-135 tanker, C-40, and the 737 commercial plane, and the Lockheed Martin-built C-17 and C-5 transports. This year, Boeing plans to use its automated technology to assess damage on the B-1 and B-52 bombers, C-130J transport, and P-8 submarine-hunting aircraft.
Belanger said the tools Boeing are developing will not replace maintainers but will better prepare them to do their jobs better. It only takes about 40 minutes for the drone to inspect an aircraft but if combined with the damage assessment software, it will help a maintenance team fine-tune their inspection plan and bring the right tools to look at a potential problem, he said.
Having the autonomous inspection capability can enhance readiness by helping maintainers quickly figure out whether an issue can be fixed on the spot or require an aircraft to be flown to a rear area for a more involved repair, he said.
Boeing plans to move “aggressively” in the next 12 to 18 months to further develop and demonstrate the automated aircraft damage assessment tools, he said.
A version of this story originally appeared in affiliate publication Defense Daily.