As drones and other airborne traffic add to traditional bird strike risks, pilots face a growing challenge during the most critical phases of flight. A new iPhone app aims to provide real-time detection of hazards before they come into view.
A Spanish artificial intelligence company is now testing a camera-based detection system designed for aviation, aiming to identify not only birds but also drones and other potential collision hazards before they are visible to the human eye.
"We developed quite a lot of know-how doing computer vision in critical and complex environments," Esteban said. "Then I thought, where can we apply this now?"
That eventually led the company toward bird strike prevention. The onboard version currently uses an iPhone mounted to the windshield with a standard suction mount. The app simultaneously accesses both the phone's telephoto and wide-angle cameras while onboard AI analyzes the live video feed for birds, drones, aircraft, and other airborne objects ahead of the aircraft.
According to Esteban, the system is designed to identify potential airborne collision risks as early as possible. "The goal here is to detect this three-pixel bird before the human eye can even recognize it," he said. Demonstrations on the company's website show the system identifying birds and other objects as small moving specks against the sky. In one test clip, the software highlights a distant object with a box outline while simultaneously labeling its classification (bird versus airplane) as it approaches, transitioning from a barely visible dot into a clearly identifiable bird.
The display overlays tracking markers and estimated distance in real time. Esteban said he was even surprised by how early the system could detect airborne objects during testing. "We were also impressed with how well it works," he said. "We were not expecting to be able to detect two or three pixels and know this is really a bird."
The company estimates the system could provide between 10 and 30 seconds of warning time for certain airborne hazards, depending on aircraft speed and conditions. Unlike radar-based systems, Aerial Guard relies on passive optical cameras and onboard processing rather than active sensors or cloud connectivity. "You can put your airplane mode on and everything works," Esteban said.
The aviation side of the project began after a Spanish flying club approached the company about adapting its ground-based detection technology for aircraft use. In Spain, pilots have expressed concern about collisions with vultures, which can pose a serious hazard to aircraft.
Pilots from the flying club have since helped the company record flight footage and test early versions of the system in flight. Esteban said some pilots were skeptical at first that a simple camera-based setup could detect birds from a moving aircraft, but early testing surprised both the pilots and the company.
The company is also developing external camera systems mounted on wing tips that would allow for stereo vision and more accurate positioning. By using two synchronized cameras, the software can estimate an object's distance and trajectory more precisely than a single-camera system.
"If you have just one camera, you see one bird, but you don't know if it's a small bird nearby or a large bird far away," he explained. "If you have two cameras, then you can compute the exact distance."
Cambrian Intelligence has also developed a ground-based version designed for airports and critical infrastructure. The ground-based version uses pairs of cameras spaced far enough apart to calculate a bird's position and trajectory in three dimensions without radar. The company says the technology can monitor bird activity, replay incidents, and analyze patterns over time.
Additional features currently under development include iPad support, integration with navigation apps, collision trajectory analysis, external camera systems, and expanded night-sky monitoring capabilities.
For now, Esteban said the company's main focus is understanding how pilots actually want to use the technology in flight and refining the system through operational feedback. "We want critics," he said. "Bad criticism is good because it allows us to improve."
Aerial Guard is available in the Apple App Store.