|
|
|
“Nonlinear Real-Time Optical Signal Processing”
by A.A. Sawchuk, T.C. Strand, and A.R. Tanguay, Jr.
June 30, 1981
The results of a four year research program in nonlinear real-time optical signal processing are described. The goal of the program was to extend fast parallel nonlinear operations to optical processing systems with large time-bandwidth and space-bandwidth products. Real-time homomorphic and logarithmic filtering by halftone nonlinear processing has been achieved. A detailed analysis of degradation due to the finite gamma and other non-ideal effects of the recording medium has been completed along with an extensive study of techniques for precompensating these effects. Further improvements in experimental work await a liquid crystal light valve (LCLV) or other real-time spatial light modulator with a sharper threshold and better uniformity and repeatability. A direct nonlinear technique of real-time parallel A/D conversion of an incoherent optical image has been developed. A detailed theoretical analysis and modeling of liquid crystal devices and associated photoconductors was completed, providing an excellent fit to experimental data obtained on several devices provided by Hughes Research Laboratories. These results have been used to make a flexible multiple light valve system (MLVS) in which the form of the nonlinearity can be varied. Variable-grating models (VGM) liquid crystal devices that perform local spatial frequency modulation as a function of the incident intensity have been developed. These devices can be used for nonlinear processing by selection and recombination of these spatial frequency components. These devices have many interesting physical effects with useful applications in both analog and numerical optical signal processing. Preliminary theoretical modeling work to explain these effects is given, and experimental implementation of binary combinatorial logic functions with VGM devices has been demonstrated. The project was a joint effort between the University of Southern California Image Processing Institute (USCIPI) and the Hughes Research Laboratories (HRL), Malibu, California.