Symbolic pattern matching for autonomous target acquisition

A symbolic pattern matching approach for imparting autonomous target acquisition capability to future imaging sensor-equipped precision-guided munitions is described. A symbolic representation of the munition's image is matched with a symbolic reference image from another sensor to find the desired target. The approach uses not only the similarity of local descriptors but also the consistency of their global configuration. The technique is designed to accommodate large differences in the viewing geometry and sensor characteristics between the two images. The method is illustrated with simulation results from FLIR imagery typical of the autonomous acquisition scenario.