Optical Wavelet Transform Processor
Abstract
In this report we outline a time-integrating hybrid-optical processor to perform the wavelet transform of two dimensional spatial or spectral data. This architecture partitions the optical/electronic workload for most general and efficient operation. This processor architecture is very general in that it allows a large class of analyzing wavelet kernels, does not require special encoding methods to accommodate the bipolar nature of the wavelet kernel, naturally accommodates the serial nature of sensor data readout, has the ability to be easily reprogrammed for different wavelet kernels, and makes most efficient use of wavelength-, spatial-, and temporal-multiplexing for fast operation. The processor is based on a triple product processor architecture which uses acousto-optic modulators and laser diodes as a means of high throughput data insertion into the system. The architecture is not limited by requiring the use of 2-D spatial light modulators and fixed masks which are the basis for many current wavelet processor concepts. We provide the result of a simulation of the optical processor to show its potential performance.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 22, 1993
- Accession Number
- ADA273625
Entities
People
- Carl C. Aleksoff
- Nikola S. Subotic
Organizations
- Environmental Research Institute of Michigan