Optical processing for landmark identification

A study of optical pattern recognition techniques, available components and airborne optical systems for use in landmark identification was conducted. A data base of imagery exhibiting multisensor, seasonal, snow and fog cover, exposure, and other differences was assembled. These were successfully processed in a scaling optical correlator using weighted matched spatial filter synthesis. Distinctive data classes were defined and a description of the data (with considerable input information and content information) emerged from this study. It has considerable merit with regard to the preprocessing needed and the image difference categories advanced. A optical pattern recognition airborne applications was developed, assembled and demontrated. It employed a laser diode light source and holographic optical elements in a new lensless matched spatial filter architecture with greatly reduced size and weight, as well as component positioning toleranced.