MULTIREFERENCE DETECTION USING ITERATIVE JOINT TRANSFORM OPERATION WITH COMPOSITE FILTERING

In this article we shall introduce an iterative joint transform operat ion for multireference detection utilizing composite intensity compens ation filtering. This technique essentially uses the normalized correl ation peak intensity values as the iterative feedback into the input a nd Fourier planes to update the weight coefficients for the reference patterns and the composite intensity compensation filter. As shown in the simulated results, this technique can detect highly similar object s which cannot normally be differentiated with the conventional joint transform correlator. The effect due to noise is also investigated, in which we have shown that the technique improves the performance of th e joint transform correlation. (C) 1994 John Wiley & Sons, Inc.