Optical Pattern Recognition
Abstract
This study provides a basic understanding of the fundamentals of optical pattern recognition systems. An emphasis on the capabilities and limitations of such systems is presented. Areas in need of further development are highlighted. Pattern recognition using optical systems can be performed based on either correlator or non-correlator systems. Recognition can be achieved by using linear processing techniques such as wavelet transforms or nonlinear processing using morphological operations or neural networks. The fundamentals of pattern recognition are presented, specifically, Fourier transformation using lenses, and the basic optical correlator architecture. Both the 4-f in-line and joint-transform correlators are described with an emphasis on the advantages and limitations of each. Morphological processors, neural- network and wavelet transform systems are introduced along with their optical implementations. A brief overview of the optical components involved in the design and fabrication of optical pattern recognition systems is discussed. Spatial filter designs used in optical correlator systems are reviewed with particular emphasis on distortion-invariant applications. Discussion of system performance criteria, and a comparison between some of the systems, are included.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1994
- Accession Number
- ADA285384
Entities
People
- G. L. Henderson
- M. A. Abushagur