Point and Shoot a Light-Field Camera, Focus Later

January 31, 2014 General, Photography 0 Comments

Point and Shoot a Light-Field Camera, Focus LaterA pioneering new camera is hoping to change the way people take pictures. The Lytro camera takes an unfocused picture that can be focused to any depth later, opening up all sorts of opportunities for professional and amateur photographers. Here's how it works and what it means for the photography industry.

How It Works

The Lytro is what is called a light-field or plenoptic camera. Originally conceived in 1908, these were not produced until quite recently. A Stanford team created and tested a working model in 2004, producing the first photographs that could be refocused after the image was captured. A commercial model for scientific and industrial use was released in 2010.

A normal camera captures the color and brightness of a two-dimensional image focused onto a flat surface. A light-field camera records all this information, but also captures the direction of the light. This obviates the need to focus the image, as focusing is primarily a function of recording the five-dimensional light-field as a two-dimensional plane. The idea of a light-field being five-dimensional in a three-dimensional space is counter intuitive, but reflects the three dimensions of space and two-dimensional representation of direction. A plenoptic camera captures four-dimensional information: At each point on a two-dimensional plane, the direction of light as measured by two angles.

As if the concept alone were not baffling enough, the workings of plenoptic cameras are extremely different from those of a typical camera. The unique technology is works thanks to the micro lens array, housed within the light-field sensor. The industrial model from 2010 contained 40,000 micro lenses; the Stanford team's 2004 model used nearly 90,000.

The array works much like the eye of an insect; each lens, positioned in a grid on the sensor, redirects an extremely small segment of the light entering the camera, providing depth information. Early designs required one lens per pixel; by placing the micro lens array farther away from the primary lens, researchers have overcome this problem along with other technical obstacles.

Beyond Neat Computer Graphics

Some applications are obvious — creating fascinating images that can be refocused by the user and the flexibility to chose the focus after taking a picture. But the light-field camera offers several other interesting possibilities both for professional photography and for practical use. The plenoptic data can be edited to produce a single photograph that is focused at all points. This technique, known as focus stacking, is already possible, but is limited because each individual focus requires a separate photograph from the same location. This both limits the number of focal levels possible and precludes moving subjects. Light-field cameras can capture moving subjects easily and allow for essentially unlimited points of focus.

The technology also has huge potential for security cameras. Images produced by such cameras are notoriously low quality, as they are focused at only one depth. If they track movement at all, they can only focus on a single moving object. Plenoptic security cameras would be able to produce clear images and potentially even create three-dimensional representations of moving objects, including suspects caught on camera.

Interest in light-field cameras is exploding, with rumors that a basic plenoptic function might be available in phone cameras as early as this year. In coming decades, the single-focus camera and photograph may well be history.

Photo credit: Wikimedia Commons.


What do you think?