What lies around the corner remains hidden to the human eye. Yet Vivek Goyal, an electrical and computer engineering professor at Boston University, wasn’t satisfied with this apparent truism of physics. As he and his team wrote in the scientific journal Nature, the ability to see around corners would have numerous advantages for safety and navigation. Autonomous vehicles would become safer, firefighters could see into burning buildings and soldiers could better assess dangerous situations. This vision might now be becoming a reality.
In 2019, the researchers managed to reconstruct a rough image from around a corner. The success is based on the assumption that scattered light rays can be put back into order and traced back to their source. Much like a mirror, even a wall also reflects rays of light; the crucial difference is in the way the light is reflected. If the light hits a shiny surface, the rays are reflected in an organized manner. But if the light hits a rough surface, they are scattered in every direction. Since the eye and brain can’t sort and process this information, the only thing perceived is a blurry reflection.
However, Goyal and his team have developed an algorithm that brings order to this chaos. For their experimental setup, the researchers placed a monitor around a corner, outside the field of view. The light from the monitor produced a blurred reflection on the wall opposite the monitor, in the team’s line of sight. Goyal took a picture of this reflection using a digital camera. He then transmitted the image to a computer. Step by step, the algorithm reconstructed the directions from which the light rays originated. The result: a pixelated yet recognizable image of the character Toad from the Mario Brothers video games, just like the one displayed on the monitor around the corner.
Goyal and his team managed to get an even sharper image when they placed an obstacle between the monitor and the wall (but out of sight of the camera). Since fewer light rays hit the wall directly with this set up, the algorithm can derive more information about the direction of the light rays and has an easier time tracking them back to their source. But the reconstructed image was not the only success of Goyal’s achievement. The professor’s experimental setup used a standard digital camera and a computer with average processing power. The use of this relatively simple and inexpensive equipment demonstrated that the technology could become commercially viable in the future—even if there are still a number of hurdles to be faced before a camera that can peek around corners is ready for mass production.