Experiments on real-time polychromatic signal detection by matched spatial filtering

Abstract Experimental results for a real-time polychromatic pattern recognition system is presented. In this technique, image recognition is performed based on both the spectral content and the shape of the target which can be a diffusely illuminated real object. The correlator employs a color encoding grating and a spatial light modulator to convert the incoherently illuminated object into spatially sampled coherent images to be processed by a multi-spectral band matched spatial filter. The advantage of this system over prior matched filtering technique is its ability to use both color and shape in recognizing real objects. Experimental results obtained by this real-time correlator indicate that the system has a high degree of color-spatial selectivity.