Eye-Inspired Camera Allows Drones to be More Autonomous

drone vision.jpg
A group of researchers from the University of Zurich and NCCR Robotics has gone to great lengths to improve how a drone sees, by trying to mimic the human eye. 

What they created was an eye-inspired camera that can both deal with high-speed motion and near-dark conditions. Which plays a crucial role in making drones become more autonomous. So, Skynet. 

The problem with current semi-autonomous drones is that they need to know their precise orientation in space at all times. This doesn't just mean GPS, either, this means spatial awareness beyond a gyroscope. It needs to be able to see whats around them and recognize what they can safely fly through or around. Buildings location may be on GPS, but not their height. 

As of now, drones use conventional cameras, which can cause motion blur when trying to use computer vision which basically renders all vision algorithms useless as they can't determine what's really there. 

Scientists have been mounting extra sensors to drones to try and help them navigate like laser scanners, but all of the extra devices can be bulky and add weight, which changes flight time, range etc. But the Swiss team believes we just have to change how the drones see, and that could change everything.

Greg Nichols from ZDNET interviewed Prof. Davide Scaramuzza, Director of the Robotics and Perception Group at UZH on how they invented the event-based camera that's bio-inspired to solve those problems:

"Event cameras are bio-inspired vision sensors that output pixel-level brightness changes instead of standard intensity frames. Crucially, the camera's retina doesn't require full light capture to generate a clear image.
"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 UZH researchers designed software to efficiently process the output from these cameras. So far they've successfully used the software to enable autonomous flight at higher speeds and in lower light than currently possible with commercial drones--at least in limited tests.
"There is still a lot of work to be done before these drones can be deployed in the real world since the event camera used for our research is an early prototype," says PhD Student Henri Rebecq. Professor Scaramuzza adds: "We think this is achievable, however, and our recent work has already demonstrated that combining a standard camera with an event-based camera improves the accuracy and reliability of the system."

Source: ZDNET