Pattern recognition with an adaptive phase-input joint transform correlator

Abstract

An adaptive phase-input joint transform correlator for real-time pattern recognition is presented. A reference image for the correlator is generated with a new iterative algorithm based on synthetic discriminant functions. The obtained reference image contains the information needed to discriminate reliably a target against known false objects and a cluttered background. Calibration look-up tables of all used opto-electronic elements are included in the design of the adaptive phase-input joint transform correlator. The resulting joint input image for the correlator is a real-valued bipolar image, which cannot be directly displayed with a conventional amplitudeonly spatial light modulator. Commonly two optical correlations and post-processing are used. We utilize a phase-only spatial light modulator in the input plane. A new phase-only joint input image is obtained by a monotonic mapping the intensity to phase information. The phase-only image is easily introduced into an optical setup. In this case we need just one correlation and no post-processing. Experimental results are provided and compared with those obtained with computer simulations.