The use of film nonlinearities in optical spatial filtering.

Optical data-processing systems that use spatial filters for detecting two-dimensional signals are usually arranged so that the data are placed in a space domain and the signals to be detected are placed in a frequency domain as holographically recorded Fourier transforms. We propose to interchange the roles of the data and the signal by introducing the data into the system as a Fourier transform hologram and by placing the signal in the space plane. Furthermore, the nonlinear behavior of the film used to record the Fourier transform hologram can be used to good advantage. These modifications combine to give a system that (1) acts as an adaptive filtering system because the noise suppression part of the matched filter is based on a specific noise sample rather than on averages of many noise samples, (2) is less sensitive to the position of the element that must be changed to perform a search for the scale or orientation of the signal, and (3) may reduce the aberrations that are sometimes caused by spatial carrier-frequency functions. Experimental results support the theoretical conclusions.