Wavelet transforms have found many practical applications in signal processing, including image processing, pattern recognition, and feature extraction. Since the wavelet transform is fundamentally given by the correlation of an input signal with a family of daughter wavelets, any real- time correlator can be used to implement the wavelet transform. Optical correlators are particularly well suited to this application, since the wavelet transform of a 1D signal produces a 2D wavelet transform. In this paper,w e discuss implementations of the wavelet transform using acousto-optic correlators and smart pixels as spatial light modulators. Smart pixels integrate both electronic processing and optical devices in a 2D array, which takes full advantage of the programmability of electronics and the parallel processing of optical devices. We describe a specific smart pixel implementation consisting of analog liquid crystal integrated on silicon 2.0-micrometers CMOS circuitry; and present experimental results of the wavelet transform implementation. An acousto-optic architecture for real-time wavelet correlators using this device will also be presented.
[1]
Francis T. S. Yu,et al.
Optical pattern recognition: architectures and techniques
,
1996,
Proc. IEEE.
[2]
Barry L. Shoop,et al.
Smart pixel-based wavelet transformation for wideband radar and sonar signal processing
,
1997,
Defense, Security, and Sensing.
[3]
Joseph N. Mait,et al.
Experimental characterization of a diffractive optical filter for use in an optoelectronic analog-to-digital converter
,
1998
.
[4]
Casimer M. DeCusatis,et al.
Wavelet transform-based image processing using acousto-optic correlators
,
1996,
Defense + Commercial Sensing.
[5]
I. Daubechies.
Orthonormal bases of compactly supported wavelets
,
1988
.
[6]
Barry L. Shoop,et al.
Experimental results from a smart pixel implementation of the wavelet transformation for signal processing
,
1998,
Defense, Security, and Sensing.