Hybrid (Optical/Electronic) Computing and Digital Optical Computing
Abstract
Hybrid optical-analog/electronic computing is performed for optical image processing, optical pattern recognition and optical solution of partial differential equations. For example, image processing designed and fabricated space variant filters to optically implement numerous space-variant transformation (e.g., Hough transform for detection of high-order parametric curves, coordinate transforms to detect rotation and scale invariant features of images, etc.). For spike removal from noisy images, one can develop a new parallel algorithm suitable for optical/analog/electronic hybrid implementation. This parallel algorithm is space-variant and performs better than the space- invariant low-pass filter and the time-consuming median filter. In digital optical computing research, develops nonlinear optical devices and studies their uses inn parallel architectures for implementation of parallel algorithms. The nonlinear optical devices are based on integrating silicon with PLZT. LPCVD techniques to deposit polycrystalline Si onto the electro-optic PLZT substrate and then recrystallizing the polysilicon using an Ar+ laser are used to produce single crystal silicon grains, to fabricated and test a 12 x 12 electrically addressed spatial light modulator (SLM) array and a 16 x 16 optically addressed SLM array. Optical parallel computing architectures, is investigated by different interconnection topologies and holographic optical elements are generated to implement fully interconnected and hypercube interconnected topologies.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 02, 1988
- Accession Number
- ADA214502
Entities
People
- Sing H. Lee
Organizations
- University of California, San Diego