Plenoptics Demystified

by admin on July 18, 2011

Despite the growing surge of interest for cameras and photography in general, little is still known about plenoptics. The term ‘plenoptics’ takes roots from words that mean ‘view’ and ‘complete’. In essence, plenoptics are specialized cameras that possess the ability to capture images and views from all viewpoints and angles. Plenoptic cameras are also commonly known as light field cameras. With an array of microlenses, plenoptic cameras are designed to capture 4D light field information. Indeed, the introduction of plenoptic cameras mean innovations in the arena of computer graphics.

A product of the brilliant minds of Edward Adelson and John Wang of MIT, plenoptic cameras were first conceptualized in 1992. The potential of depth estimation and synthetic refocusing became the sources of inspiration for such a brilliant innovation. Adelson and¬†Wang’s version also went through enhancements introduced by Ren Ng and colleagues from Stanford University in 2005.

In a single lens stereo design, a camera can capture an image through a conventional lens. It is also possible to capture an image with an eccentric lens. The image a plenoptic camera can take a shot of largely depends on the displacement of the aperture. For instance, left displacement of the aperture indicates that the plenoptic camera can take a shot of a near object with left displacement and of a far object with right displacement. Simply said, the aperture displacement will be copied by a near object and will be opposed by a far object. With that, a plenoptic camera indeed possesses the needed features to be an efficient depth estimator.

In estimating depth with a single lens, simple geometric analysis is integrated. However, one snapshot is not enough for a complete or comprehensive depth estimation. If you want to have a good estimation of depth, you need to collect numerous snapshots. Indeed, with plenoptic camera, it is possible to acquire everything at once.

The design of a plenoptic camera resembles the image formation in an ordinary camera. A microlens array is inserted before the sensor to make the capture of snapshots possible. Microlens is being used to capture the distribution of directional lighting arriving at each of the available sensor locations. This information about the direction is significant because it provides the user the opportunity to rearrange the pixels according to his or her own preferred order. The rearrangement of the pixels also lead to the rearrangement of all the rays associated with each of the pixels. With the possibility of the rearrangement of the rays, the possibility of synthetic refocusing becomes even more salient.

Synthetic refocusing enables the camera to capture a snapshot once and produce numerous shots focused at varied depths. When you resort the pixels, you also resort the rays. When you are already resorting the rays, you are also refocusing.

Plenoptic cameras have actually been intended to minimize the correspondence problem in stereo matching. Thus, the microlens is situated at the focal plane of the main lens of the camera for strategic purposes. Slightly behind the microlenses, you can also see the image sensor.

Now, a lot of companies are considering investing in plenoptic camera and introducing it to the market. The first company to pave the way for plenoptic cameras to infiltrate the local photography market is the German company, Raytrix. Its plenoptic camera offers the user the opportunity to identify focus points in post processing and use one sensor only in capturing snapshots of 3D images. You can choose from Raytrix’s list of plenoptic cameras, from R5 to R11. The R5 camera is considered the ‘entry level’ variety since its shooting capacity is limited to 1 megapixel only. If you want a camera that can shoot at 3 megapixels, the high-end R11 Raytrix plenoptic camera would be better suited for your preference and need. What makes it even more amazing is the opportunity to convert any available digital camera into a plenoptic camera. How is this possible? You can just simply create a lens array for it and it will only take 6 to 8 weeks until your digital camera becomes a plenoptic camera.

The company owned by an alumnus of Stanford University, Ren Ng, will also be introducing their own plenoptic camera to the photography market. With a single exposure and only one camera, you can take photos and even change their point of focus afterwards. Being a light-field or plenoptic camera, Ng’s camera uses an array of microlenses over the sensor. The sensor sits slightly behind the focal plane of the camera where the lenticular array also sits on. The light is focused by the lens or film plane on the focal plane.

Ng’s camera also captures the direction of the light, in addition to the usual color and intensity. This feature makes it possible to create astounding photographic results using some heavy processing techniques. Software also replaces most of the camera’s precision software.

Indeed, with these innovative features, we have yet to see the promise of the new technology. Plenoptic cameras will certainly turn your world around. Plenoptic cameras will certainly introduce a new way of looking at things, literally and figuratively.




{ Comments on this entry are closed }