CMU researchers use computer vision to see around corners

by

in

Future autonomous vehicle as well as other machine intelligence systems might not need line-of-sight to gather incredibly detailed picture info: New research from Carnegie Mellon University, the University of Toronto and University College London has invented a method for “watching around corners. ”

The system employs special sources of lighting, combined with sensors and computer vision processing to efficiently infer or rebuild extremely detailed vision, much more detailed than has been possible previously, without having photographed it otherwise ‘viewed’ it straight.

There are a few limitations — so much, investigators working on the job have only been able to use this procedure effectively for “comparatively smaller places,” based on CMU Robotics Institute Professor Srinivasa Narasimhan.

That restriction may be mitigated by applying this technique alongside others utilized in the field of non-line-of-site (or NLOS) computer vision research. Some such processes are already in use on the marketplace, including the way Tesla’s Autopilot system (along with other driver-assist technologies) uses mirrored or bounced radar signals to find out around the cars immediately in front of this Tesla automobile.

The technique used in this study is really like what happens in a LiDAR system used in many autonomous vehicle systems (though Tesla famously eschews utilization of laser-based vision systems in its own technology pile ). CMU and its partner institutions use ultrafast laser light inside their own system, bouncing it off a wall to lighting an object hidden around a corner.

Sensors then capture the reflected light when it bounces straight right back, and investigators measure and calculate how long it took for the reflected light to come back to the point of source. Taking quite a few measurements, and using information about the goal object’so called geometry, the group managed to reconstruct the objects with remarkable detail and accuracy. Their method was so effective that it even works via semi-osbscruingf substances, including heavy paper — yet another big advantage when it has to do with its potential to be used in surroundings sensors that operate in real-world ailments.

Buy Tickets for every event – Sports, Concerts, Festivals and more buy tickets

Discover more from Teslas Only

Subscribe now to keep reading and get access to the full archive.

Continue reading