Hybrid pattern recognition by features extracted from object patterns and Fraunhofer diffraction patterns

A hybrid method of pattern recognition is proposed. In the method, an object image is observed by a TV camera and then the shift- and rotation-invariant features of the image are extracted by digital image processing. At the same time, the Fourier spectrum of the object image is obtained by a coherent optical system and the shift- and rotation-invariant features are extracted from a characteristic function derived from the spectrum. Classification of the object is carried out by a small-scale digital processor, using both features obtained by digital and analog image processing.