Scientists Use Lasers to See Around Corners
The mannequin, shown here, is not in the camera's line of sight.
CREDIT: Christopher Barsi and Andreas Velten, MIT Media Lab
Using a combination of ultra-fast lasers and powerful software, scientists have devised a way to create 3-D images of objects hidden around corners. The system could allow rescue workers to one day peer into places that are difficult or hazardous to enter.
Laser scanners are now often used to capture 3-D images of items. These work by bouncing pulses of light off targets — since light travels at a constant speed, measuring the amount of time it takes for these pulses to return tells you the distance they traveled, data which can be used to build what an object looks like in three dimensions.
Laser scanners generally look at objects directly in front of them. In principle, if they used a highly reflective surface like a mirror, they could readily see items obscured from their line of sight, such as something hidden behind a corner. However, if only poorly reflective surfaces are available, such as most walls, then light gets scattered diffusely and the image gets lost.
Now scientists reveal they can see objects around corners even using only typical walls.
"Most imaging systems are modeled after the human eye, but we are now able to use light in completely different ways and can come up with these surprising new abilities," researcher Andreas Velten, an imaging scientist at MIT, told InnovationNewsDaily.
The key is using an ultra-fast pulsed laser, one with infrared pulses lasting just 50 femtoseconds long, a femtosecond being one-millionth of one-billionth of a second. The camera they used works quickly as well, capturing images every 15 picoseconds, with a picosecond being one-millionth of one-millionth of a second.
"The length of the pulse translates into the resolution of our image," Velten said. "One picosecond at the speed of light is a little less than 1 millimeter. This is about the resolution we can get."
The researchers fire laser pulses off a white wall toward a hidden object. Some of the scattered light reaches the object, reflects back, and gets scattered by the wall again before reaching the scanner.
By tracking the return time of the scattered light, computer algorithms could untangle all the data hidden in this scattered light to reconstruct the full 3-D shape of the hidden object — for instance, an 8-inch-tall (20-centimeter-tall) mannequin.
Applications for this research could include "anywhere where you want to look into a space that is beyond your direct line of sight," Velten said. "Search-and-rescue operations and disaster response are good examples — learning more about the sites of chemical or nuclear accidents or fires is immensely important. Even when you are able to bring a camera on a robot into the scene, it turns out to be very hard for the robot operator to quickly figure out what is going on from the limited camera image."
Industrial inspection involving peering inside machinery with moving parts is another potential application. "You also may one day have one of these in your car, so it can warn you when there is an obstacle around the corner," Velten said.
Future research could improve the system so that it can work on room-sized scenes, or testing if it might be useful in medical applications inside the body, Velten said.
The scientists detailed their findings online March 20 in the journal Nature Communications.