Eye-inspired camera allows drones to fly in low light conditions Friday, 22 September 2017

Researchers from the University of Zurich (UZH) say drones of the future could be able to see better in low light conditions, allowing them to perform fast, agile manoeuvres and fly rescue and search missions at dusk or dawn.

To fly safely, drones need to know their precise position and orientation in space at all times. Commercial drones do this using GPS, but that requires them to be outdoors, and GPS can at times be unreliable, especially in urban environments.

Additionally, drones currently use conventional cameras that require a decent amount of light to work optimally. In dim conditions, images from the cameras are motion-blurred and cannot be used by computer vision algorithms, necessitating a reduction in the speed of the drones.

To solve this problem, professional drones use sensors that are elaborate, expensive, and bulky, such as laser scanners.

The alternative approach by the researchers, in cooperation with the Swiss research consortium NCCR Robotics, uses an eye-inspired camera that can cope with high-speed motion, and is able to see in the dark much more effectively than conventional cameras. The team has already taught drones to use their onboard cameras to infer their position and orientation in space, allowing the drones to fly in a wide range of conditions

"This research is the first of its kind in the fields of artificial intelligence and robotics, and will soon enable drones to fly autonomously and faster than ever, including in low-light environments," says Prof. Davide Scaramuzza, Director of the Robotics and Perception Group at UZH.

The cameras, called 'event cameras', do not need to capture full light on the bio-inspired retina in order to have a clear picture. They only report changes in brightness for each pixel, ensuring perfectly sharp vision even during fast motion or in low-light environments.

The researchers have also designed new software able to efficiently process the output from such cameras.

Drones equipped with an event camera and the software designed by the researchers could assist search and rescue teams in scenarios where conventional drones would be of no use – for example on missions at dusk or dawn or when there is too little light for normal cameras to work. They would also be able to fly faster in disaster areas, where time is critical in saving survivors.

Before the drone and camera system can be rolled out commercially, the researchers will have to prove that the software also works reliably outdoors. Recent findings have also demonstrated that combining a standard camera with an event-based camera improves the accuracy and reliability of the system.