Now You Know

MIT Speeds Up The Single-Pixel Camera

Courtesy of the researchers
Courtesy of the researchers

A single-pixel camera, as its name implies, uses a a single light sensor rather than the millions of light sensors usually found in a conventional camera. Researchers at Rice University built such a camera in 2006 that could produce 2-D images using a single light sensor.

The single-pixel camera captures an entire image by taking thousands of exposures using the single light sensor, in effect compressing them using a technique called “compressive sensing.” The reason the single-pixel camera can make do with one light sensor is that the light that strikes it is patterned, using either a filter or by bouncing the returning light off of an array of tiny micromirrors, some of which are aimed at the light sensor and some of which aren’t. The sensor makes only a single measurement — the cumulative intensity of the incoming light. But if it repeats the measurement enough times, and if the light has a different pattern each time, software can deduce the intensities of the light reflected from individual points in the scene, and thus recreate the scene.

But compressive sensing can be quite slow: it takes about five minutes (and thousands of exposures) to take a reasonably clear picture with Rice’s prototype camera. The MIT researchers (Guy Satat, Matthew Tancik and Ramesh Raskar) have devised a new technique that significantly speeds up that image acquisition process by about 50x and getting the number of exposures required down from thousands to dozens.

In the new technique, a filter with a particular pattern is placed in front of the light source that illuminates the scene. The patterned light is reflected from the scene onto the detector. The detector is ultrafast and time tags photons as they arrive. Photons from sources close to the detector arrive before photons from further sources. Different patterns are used and the resulting measurements are then used to computationally reconstruct the scene. This time resolved detector requires the use of much fewer patterns than traditional single-pixel cameras. By optimizing the patterns, the scientists believe that even less of them may be required.

Interestingly, a single-pixel camera does not require a lens to capture an image, potentially opening up new prospects for the design of imaging systems and making them useful in harsh environments or in applications that use wavelengths of light outside the visible spectrum.