CHAMPAIGN, Ill. -- Using principles gleaned from radio astronomy and medical X-ray tomography, researchers at the University of Illinois have assembled an optical system that produces three-dimensional reconstructions of objects without using a lens.
"One big advantage of designing a lensless 3-D camera is that the resulting optical system has an infinite depth of field," said Ronald Stack, a research engineer with the U. of I.'s Beckman Institute for Advanced Science and Technology. "That means the object will always be in focus, which simplifies the amount of data processing that is required."
Instead of a lens, the camera system uses a series of beam splitters and folding mirrors to capture a sequence of "snapshots" as an object is slowly rotated in front of the aperture. The images are detected with a CCD (charge-coupled device).
"The snapshots are not like ordinary photographs, however," said David Munson Jr., a professor of electrical and computer engineering and a researcher at the Beckman Institute. "The camera collects data on interference patterns in much the same way as does a radio telescope. These two-dimensional data frames are then processed to form a 3-D representation of the object, which can be displayed on a computer screen or in a virtual reality environment."
To produce the 3-D image, the researchers rely on a reconstruction algorithm used in cone-beam tomography -- a relatively new medical X-ray photography technique that produces 3-D images of a patient, instead of the more familiar cross-sectional slices.
"With our camera, we can reconstruct 3-D objects purely from physical principles, field analysis and number crunching," Munson said. "Because our system is not based on conventional computer vision and image-processing techniques, we incorporate no tricks, heuristics or data manipulation."
Such physical optics techniques may ultimately benefit areas such as microscopy and machine vision by providing 3-D reconstructions of superior resolution.
"In a lot of current 3-D visualization activities, people need a quick and reliable way to capture 3-D information about the optical world to put into their database," Stack said. "Our camera system can also help do that, because it produces a numerical 3-D representation of the object."
In addition to Stack and Munson, the research group included team leader David Brady, a professor of electrical and computer engineering and a researcher at the Beckman Institute; Daniel Marks, a graduate research assistant at the Beckman Institute; and Rachael Brady, an expert on scientific visualization at the university's National Center for Supercomputer Applications.
The researchers described the lensless imaging system in the June 25 issue of the journal Science.
Cite This Page: