Nothing short of potentially heralding a new age of camera design? As we all know there is no camera without a lens, a physical necessity carved in stone since the invention of photography. Lenses today are near perfect, we think, and we’re willing to spend top dollars on them. Now a team of U.S. engineers in Texas is about to change that by building a camera that can make a sharp image with no lens at all.
Their so-called FlatCam is based on a standard imaging sensor paired with a mask and decoding software. They “challenge the 3-dimensionality, the chunkiness of camera designs by a new approach that can make cameras very, very flat,” says Richard Baraniuk, professor of computer and electrical engineering at Houston’s Rice University. So that cameras can be very thin, very flexible and they can be used in a lot of applications where lens-bases cameras just can’t be used.
First, they lay a very thin mask right on top of the sensor, and then the image is extracted by computational reconstruction. Computer algorithms do the task of focusing and reconstruction. That’s right, just replace lenses with masks and computation.
These are early days, says Baraniuk. Right now the 512 x 512 pixel images FlatCam can produce are about as good as the first conventional digital cameras, but he has no doubt quality will get better very soon. He expects people will come up with ways to use the new camera that he can’t even imagine yet.
Something he could imagine: “Making, for example, wallpaper that you could paper over a wall to create an extremely massive camera,” he says. Using an entire wall as a camera means you could see absolutely everything in the room, including things that would be out of sight or distorted in a single, smaller camera.
Or build a cylindrical camera, put something in the middle of the cylinder, and you can take pictures of it from every angle at the same time. Or just imagine the potential for industrial or medical devices — devices that could be flexible or curved, frankly impossible with a lens.
To design FlatCam, Baraniuk and his colleagues looked to the past for inspiration — back to really the very first cameras, pinhole cameras. Pinhole cameras have been here for quite a while. According to some scholars, they were first described by the Chinese philosopher Mo Ti around 400 B.C. These cameras later came to be known as the camera obscura. Light enters through a small hole into a darkened space, sometimes as large as a room, and an image of what the hole was pointed at appears on a screen.
With the invention of film and later photo sensors, it became possible to capture that screen image.
“The disadvantage of a pinhole camera is that while it provides the simplicity of interpretation of the image, it lets very little light through,” Baraniuk says. “So it’s very inefficient that way. It’s a pinhole after all.”
So he and his colleagues decided to build a camera based not around a single pinhole like you might build for a science fair project, “but one with literally millions of pinholes,” Baraniuk says.
But there’s a problem. A million pinholes produce a million images, all smeared on top of one another.
How to separate them?
Says Ashok Veeraraghavan, assistant professor of electrical and computer engineering at Rice, the answer is computation. Lots of digital cameras use computation to improve the quality of the image created by the lens in a traditional camera: “The entire community of computational imaging has started recognizing that computation can not only be used to improve images that have been captured by the earlier camera designs, but actually change these camera designs in radical fashions.”
Such as a camera made from a million pinholes.
An age-old concept to try and reimagine the camera as we know it… Wild thought that tomorrow’s cameras might in fact based on the very first cameras every used.