New optical methods for signal processing.
Item
-
Title
-
New optical methods for signal processing.
-
Identifier
-
AAI9304751
-
identifier
-
9304751
-
Creator
-
Zhang, Yan.
-
Contributor
-
Adviser: Yao Li
-
Date
-
1992
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Electronics and Electrical | Physics, Optics
-
Abstract
-
This doctoral thesis studies the optical implementations of various new algorithms and methods for large bandwidth signal and image processing. Among the schemes to be studied are the long data stream convolution/correlation, the Gabor and the wavelet transforms, and their applications to system failure prediction, dense target signal processing and image coding.;Based on the Chinese remainder theorem, optically implementable algorithms are described, which convert the convolution/correlation of long data streams to relatively small scale linear operations such as a group of short-term vector-matrix multiplications or short-term convolutions/correlations. The proposed algorithms can be realized by using the existing optical analog data processors. Simulations were performed to prove their validity. Technical problems and fundamental limitations of the above schemes are studied.;Following the consideration of the above time domain operations, signal's representations in joint time-frequency (scale) domain are then considered. An opto-electronic Gabor coefficient processor is designed to perform the Gabor transform on short one-dimensional (1-D) signals in real-time. Some experimental results are presented to confirm the operational principle of the system. As an application of this processor, Gabor transform based transient signal detection is studied. Other schemes for implementing Gabor transform of long 1-D signals based on the long data stream convolver, and 2-D signals are also investigated. Following the study of the Gabor transform, the newly suggested wavelet transform is considered for its optical implementation. Using commercially available opto-electronic components, an optical wavelet processor is designed and built to perform the wavelet transforms on short 1-D signals in real-time. As an extension, architectures for 2-D optical wavelet transform are also described and computer simulated with the consideration of their technical problems of optical implementation. Applications of the wavelet transforms to system failure prediction, dense target signal processing and image coding are briefly explained. Schemes for optical wavelet transform of long 1-D signals are also studied.;The goal of this thesis is to further our knowledge and understanding on how the free-space optical processing methods could help developing efficient signal and image processing tools and what the potential applications and limitations of these optical methods are.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
