Researchers at MIT Media Lab have developed a $500 "nano-camera" that
can operate at the speed of light. According to the researchers,
potential applications of the 3D camera include collision-avoidance,
gesture-recognition, medical imaging, motion-tracking and interactive
gaming.
The team which developed the inexpensive "nano-camera"
comprises Ramesh Raskar, Achuta Kadambi, Refael Whyte, Ayush Bhandari,
and Christopher Barsi at MIT, and Adrian Dorrington and Lee Streeter
from the University of Waikato in New Zealand.
The nano-camera
uses the "Time of Flight" method to measure scenes, a method also used
by Microsoft for its new Kinect sensor that ships with the Xbox One.
With this Time of Flight, the location of objects is calculated by how long it
takes for transmitted light to reflect off a surface and return to the
sensor. However, unlike conventional Time of Flight cameras, the new
camera will produce accurate measurements even in fog or rain, and can
also correctly locate translucent objects.
What makes it difficult
for conventional Time of Flight cameras to accurately measure the
distance light has travelled (in a changing environment, along
semi-transparent surfaces, or to an object in motion) is the creation of
multiple reflections, which smear the original signal's reflection
before being collected by the sensor, resulting in inaccurate data. To
avoid the problem, the team used an encoding technique that is used in
the telecommunications industry.
Raskar, associate professor of
media arts and sciences, and leader of the Camera Culture group at the
Media Lab, explains the new method: "We use a new method that allows us
to encode information in time. So when the data comes back, we can do
calculations that are very common in the telecommunications world, to
estimate different distances from the single signal."
Kadambi
adds: "By solving the multipath problem, essentially just by changing
the code, we are able to unmix the light paths and therefore visualize
light moving across the scene."
Raskar's group had unveiled a
trillion-frame-per-second "femto-camera" in 2011, which costs roughly
$500,000 to build. The technique scans a scene with a femtosecond (one
quadrillionth of a second) impulse of light, and then uses extremely
expensive laboratory-grade optical equipment to capture an image each
time.
The team's "nano-camera" instead measures the scene with a
continuous-wave signal, oscillating at nanosecond periods (one
thousand-millionth of a second). According to the researchers, this
means the nano-camera can reach a "time resolution" (size of interval
between images) within one order of magnitude of femtophotography, while
costing just $500. They call the technique nanophotography, which is
claimed to deliver similar results to femtography, with marginally lower
quality, and a fraction of the cost.
