MIT camera grabs 3D images 'round corners (images)
Researchers in MIT's Media Lab have built a laser-powered camera that sees around corners at 1 trillion frames per second. Out of sight no longer means out of mind.
Femtosecond camera can peek around a corner
Photography is all about light. Typically, a lens is used to focus the light reflected from an object, striking a light-sensitive surface inside a camera to create an image recognizable to the human eye.
Until now, that light required a direct line of sight, but at MIT Media Lab's Camera Culture group, Ramesh Raskar and Andreas Velten have devised a new way of capturing light bouncing from around a corner -- a femtophotography. In traditional photography, the speed of light is infinite and does not play a role. The Media Lab's femtocamera, however, has a finite amount of time light takes to travel from one surface to another, which provides useful information.
Using a beam-splitter and what they call a femtosecond laser to send out a laser pulse that lasts less than one-trillionth of a second, the light returning from the scene is collected by a camera at the equivalent of close to 1 trillion frames per second, analyzed, and decoded to record the location, and ultimately shape of an object outside of the direct line of sight.
Until now, that light required a direct line of sight, but at MIT Media Lab's Camera Culture group, Ramesh Raskar and Andreas Velten have devised a new way of capturing light bouncing from around a corner -- a femtophotography. In traditional photography, the speed of light is infinite and does not play a role. The Media Lab's femtocamera, however, has a finite amount of time light takes to travel from one surface to another, which provides useful information.
Using a beam-splitter and what they call a femtosecond laser to send out a laser pulse that lasts less than one-trillionth of a second, the light returning from the scene is collected by a camera at the equivalent of close to 1 trillion frames per second, analyzed, and decoded to record the location, and ultimately shape of an object outside of the direct line of sight.
Determining shape and location
To create an image of the hidden object, the camera aims its laser at a nearby (nonmirrored) wall or door, bouncing the light and returning it to the sensors, which determine the shape and location of the object based on the distance and time the laser traveled.
Currently, the resolution is low, and the camera is limited to capturing centimeter-size details at a distance of a few meters. That means it can only see relatively large objects.
The laser bounces light over and over again at different angles, and eventually is able to record details of an object's position by analyzing these measurements.
Currently, the resolution is low, and the camera is limited to capturing centimeter-size details at a distance of a few meters. That means it can only see relatively large objects.
The laser bounces light over and over again at different angles, and eventually is able to record details of an object's position by analyzing these measurements.
Femtosecond laser meets picosecond-accurate camera
The imaging-camera prototype consists of a femtosecond laser illumination coupled with a picosecond-accurate camera and an inversion algorithm, resulting in what Raskar, an associate professor at the MIT Media Lab, calls a "space time impulse response" (STIR) of the scene that allows the camera to reconstruct the hidden surface.
Mirrors need not apply
The principles behind this futuristic camera system are similar to how a periscope works, reflecting an image at an angle into the viewer's eye. But this digital imaging system doesn't need mirrors. Instead, it uses ordinary walls, doors, or floors to bounce or reflect bursts of laser light.
From 'raw streak' to heat map
Many "raw streak" images from a pass of the laser, like the one seen here, are used to create a heat map of depth measurements, which is then further decoded to a visual representation of the hidden object. Light travels at about 1 foot/nanosecond and by sampling the light at picosecond resolution, the researchers at MIT say they can estimate shapes with centimeter accuracy.
You can run, but you can't hide
Once the split beam laser was decoded, MIT was able to look around corners and "see" a three-dimensional object.
Search and rescue applications
Potential applications for this imaging system include search and rescue efforts, in which clear lines of sight may not be possible.
Safer cars, smarter surgery
Transient imaging also has significant potential benefits in collision avoidance for advanced autonomous cars or for robots. There could be medical applications as well -- for instance, allowing endoscopes to view around obstacles inside the human body.
More Galleries
My Favorite Shots From the Galaxy S24 Ultra's Camera
My Favorite Shots From the Galaxy S24 Ultra's Camera
20 Photos
Honor's Magic V2 Foldable Is Lighter Than Samsung's Galaxy S24 Ultra
Honor's Magic V2 Foldable Is Lighter Than Samsung's Galaxy S24 Ultra
10 Photos
The Samsung Galaxy S24 and S24 Plus Looks Sweet in Aluminum
The Samsung Galaxy S24 and S24 Plus Looks Sweet in Aluminum
23 Photos
Samsung's Galaxy S24 Ultra Now Has a Titanium Design
Samsung's Galaxy S24 Ultra Now Has a Titanium Design
23 Photos
I Took 600+ Photos With the iPhone 15 Pro and Pro Max. Look at My Favorites
I Took 600+ Photos With the iPhone 15 Pro and Pro Max. Look at My Favorites
34 Photos
17 Hidden iOS 17 Features You Should Definitely Know About
17 Hidden iOS 17 Features You Should Definitely Know About
18 Photos
AI or Not AI: Can You Spot the Real Photos?