Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory, or CSAIL for short, have developed a surprisingly simple method that allows cameras to “see” around corners.
The technique, dubbed CornerCameras, relies on the disruptions in light that objects cause as they move. By magnifying standard video footage, the lighting changes can be differentiated and used to construct a series of one-dimensional images that reveals information about an object over time.
Fine details like an object’s shape or texture can’t be discerned but you can gather information regarding the relative speed and position of an object and in some cases, even its color. Best yet, the technique works in real-time and under subpar lighting conditions (outside in the rain, for example).
As for practical applications, researchers believe the technology could one day be embedded into smartphones and help self-driving cars see around corners to spot pedestrians or other vehicles.
Getting to that point will take some work, however, as a laptop is currently needed to handle processing duties. What’s more, the system in its current iteration requires a stationary camera although the team is refining it to work with motion.
Found is a TechSpot feature where we share clever, funny or otherwise interesting stuff from around the web.