Real-time optical image processing techniques

Nonlinear real-time optical processing on spatial pulse frequency modulation has been pursued through the analysis, design, and fabrication of pulse frequency modulated halftone screens and the modification of micro-channel spatial light modulators (MSLMs). Micro-channel spatial light modulators are modified via the Fabry-Perot method to achieve the high gamma operation required for non-linear operation. Real-time nonlinear processing was performed using the halftone screen and MSLM. The experiments showed the effectiveness of the thresholding and also showed the needs of higher SBP for image processing. The Hughes LCLV has been characterized and found to yield high gamma (about 1.7) when operated in low frequency and low bias mode. Cascading of two LCLVs should also provide enough gamma for nonlinear processing. In this case, the SBP of the LCLV is sufficient but the uniformity of the LCLV needs improvement. These include image correlation, computer generation of holograms, pseudo-color image encoding for image enhancement, and associative-retrieval in neural processing. The discovery of the only known optical method for dynamic range compression of an input image in real-time by using GaAs photorefractive crystals is reported. Finally, a new architecture for non-linear multiple sensory, neural processing has been suggested.