DESIGN & PRODUCTS OPTOELECTRONICS
Event-camera equipped drones get milliseconds reaction times
By Julien Happich
Leveraging so-called event cameras with a proprietary
moving-obstacle detection algorithm, researchers from
the University of Zurich and NCCR Robotics have reduced
a drone‘s reaction time from tens of milliseconds to only 3.5 milliseconds
in order to detect and avoid fast-moving objects.
Sequence of an avoidance manoeuver.
On the left side, the 3D volume of pixel events
accumulated during an arbitrary time window
of 10ms. On the right side, the same events
after ego-motion compensation, backprojected
on the image plane.
The findings published in the Science Robotics journal under
the title „Dynamic Obstacle Avoidance for Quadrotors with
Event Cameras“ report an overall detection and reaction latency
(to initiate an obstacle avoidance manoeuver) of only 3.5 milliseconds,
sufficiently low for the drone to avoid even fast-moving
obstacles such as a basket ball being thrown at it.
Event cameras are bio-inspired sensors that measures perpixel
brightness changes, asynchronously, capturing a stream of
events at microsecond resolution instead of having to generate
full image frames as conventional cameras do.
But standard vision algorithms that analyse image changes
across full frames cannot be applied to a stream of asynchronous
events, and the researchers had to develop novel algorithms
to exploit the temporal information contained in the event
stream to distinguish between
static and dynamic objects. Their
moving-obstacle detection algorithm
works by collecting events
during a short-time sliding window,
it then compensates for the motion
of the drone (ego-motion) during
that time. Tightly integrated
with the drone‘s inertial measurement
unit data, this ego-motion
compensation approach enables
even a moving sensor to remove
the pixel events generated by the
static parts of the drone‘s environment,
leaving dynamic objects to be easily identified.
With an overall latency of only 3.5ms, the algorithm is sufficiently
fast to run in real time on a small onboard computer.
Object detection was successfully tested both indoor and outdoor
as demonstrated in a video where a quadcopter is shown
executing various collision-avoidance manoeuvers in front of
a ball. More than 90% of the time, the drone avoided the ball
(thrown from a 3-meter distance at 10m/s). When the drone
“knew” the size of the object in advance, one camera was
enough, but when it had to face objects of varying size, two
cameras were used to enable stereoscopic vision.
Ultra-thin flat lens has extreme depth of focus
TBy Julien Happich hanks to an innovative design methodology,
solving a non-linear inverse
problem via parameters optimization,
a team of researchers from the University
of Utah has designed an ultra-thin flat lens
focused in a large focal range spanning
several metres. In a paper titled “Extremedepth
of-focus imaging with a flat lens”
published in the OSA Optica journal, the
researchers report a very thin lens that
does not require refocusing even to capture
different object images separated by large
distances, removing the need for refocusing
mechanisms and drastically simplifying
Schematic of a multi-level diffractive
lens (MDL) that exhibits extreme-depthof
focus (ExDOF) imaging.
imaging systems.
The so-called multi-level diffractive lens (MDL) described
in the paper enhances the depth of focus by over 4 orders of
magnitude compared to traditional fixed lenses. In experiments
using such a lens, the researchers were able to maintain focus
for objects separated by as large a distance as 6 metres. When
illuminated by collimated light at 0.85m, a 1.8mm diametre
multi-level diffractive lens shaped with a distribution of concentric
rings of varying thickness up to 2.6μm maximum, produced
a beam which remained in focus from 5 to 1200 mm.
The researchers
have founded
a startup Oblate
Optics in San Diego
to commercialize
planar optics with
large bandwidths
and a large depth of field. They are seeking collaboration with
strategic partners.
Optical micrograph of the fabricated
MDL, with the inset showing a magnified
view of the centre of the lens. Focal
range is 5 to 1200mm, aperture is 1.8
mm.
44 News April 2020 @eeNewsEurope www.eenewseurope.com
/eenewseurope
/www.eenewseurope.com
