A team of researchers at Rice University are developing a thin, lensless camera called FlatCam, which replaces the traditional lens with a grid-like coded mask essentially made of multiple pinholes positioned directly over the sensor. From there, raw data from the camera – which looks nothing like an actual image – is sent to a computer where an algorithm demultiplexes the raw sensor measurements to reconstruct a photo.
 |
FlatCam was developed by Richard Baraniuk and Ashok Veeraraghavan, and while it isn’t likely to replace traditional cameras, FlatCam could find use in a variety of applications where present imaging technology isn’t suitable. Due to its lensless design, FlatCam is thinner than a dime and can be ‘fabricated like microchips,’ decreasing costs.
Offering an example of possible future applications, Baraniuk explained, ‘We can make curved cameras, or wallpaper that’s actually a camera. You can have a camera on your credit card or a camera in an ultrathin tablet computer.’ The team’s paper details more potential applications, including surveillance, foldable cameras and wearable devices. They’ve even created a couple of video clips using the prototype camera.
In its present form, FlatCam produces 512 x 512 pixel images; however, the team expects better algorithms and improved manufacturing will one day result in higher resolutions. One of the salient issues mentioned in the paper surrounds noise: a linear demultiplexing system invariably adds noise from all the mathematical operations, with higher spatial frequencies experiencing even more ‘noise amplification’ that ultimately limits resolving ability. These are issues the researchers are actively working on.
The researchers recently talked about FlatCam on NPR’s All Things Considered; the episode can be streamed here.
Via: Rice University
Articles: Digital Photography Review (dpreview.com)