In an alternative to the optical image-compression method of the preceding article, the Fourier transform of the input image would be formed on the output plane by white-light holography, instead of by laser holography. The principal advantage of the alternative method would be decreased power consumption: A state-of-the-art liquid-crystal spatial light modulator needed to implement the method of the preceding article consumes about 10 W of operating power, and the liquid crystals must be maintained at a temperature near 25 °C. On the other hand, an image detector of the active-pixel-sensor (APS) type, needed to acquire the Fourier-transform image in both the method of the preceding article and in the alternative method, consumes only about 50 mW. Because the spatial light modulator would not be needed in the alternative method, the power consumption of the image-compression system could be greatly reduced.

Two White-Light Fourier-Transform Images with different magnifications would be made to interfere with each other to obtain a white light hologram on the detector plane.

In the alternative method (see figure), the input image could be formed in sunlight or any other suitable incoherent light. A lens would act as an optical Fourier transformer, but in this case, two slightly different white-light Fourier-transform images would be made to interfere with each other to form a white-light hologram. An interferometric loop would extract the relevant phase information. The particular configuration of the interferometric loop, originally developed for white-light holography, would simultaneously optimize the spatial and temporal overlap of the input intensity pattern, as needed to obtain a measurable contrast for white-light interference. Different magnifications would be used along the clockwise and counterclockwise circuits of the interferometer, in order to make each point in the image interfere with itself; this design would eliminate the need for a spatial light modulator and lasers to obtain the coherence needed to generate a hologram. The design problem then becomes one of matching the contrast ratio of the white-light hologram with the dynamic range and offset of the image detector.

The input image would be presented at the left focal plane of the Fourier-transform lens, which would have a focal length f1. The Fourier-transform image would be formed at a distance f1 to the right of this lens, at the entrance to the interferometric loop. Within the interferometric loop, two lenses with differing focal lengths of f2 and f3, respectively, would refocus the Fourier-transform image onto the image detector at two different magnifications. The magnification of the clockwise circuit would be f3/f2, while that of the counterclockwise circuit would be f2/f3. Components of the light that came from the same point in the Fourier-transform plane would interfere on the detector plane and thereby provide phase information. Light coming from different points in the Fourier-transform plane would be superimposed incoherently on the detector plane. As in the method of the preceding article, the low-spatial-frequency image information would be concentrated about the optical axis on the detector plane.

The major drawback of white-light holography is that the coherently interfering waves comprise only a small fraction of the total light, so that interferograms are unavoidably superimposed on bright backgrounds. The use of an electronic (as opposed to a photographic-film) image detector would make it possible to process the image information electronically to remove the bright background. Implementation would involve a formidable challenge in that it has been estimated that 50 dB of dynamic range would be needed to eliminate the background signal while an additional 30 dB of dynamic range would be needed to achieve 10-bit precision in pixel readouts. One could use a narrow-band filter to reduce the brightness of the background light and increase the coherence length of the light to obtain more margin for design specifications.

This work was done by Deborah Jackson of Caltech for NASA's Jet Propulsion Laboratory. For further information, access the Technical Support Package (TSP) free on-line at  under the Physical Sciences category.


This Brief includes a Technical Support Package (TSP).
High-Speed Optical Image Compression at Low Power

(reference NPO-20639) is currently available for download from the TSP library.

Don't have an account? Sign up here.