Proceedings PaperComparison of nonlinearly transformed filters and nonlinear joint transform correlators for optical pattern recognition
|Format||Member Price||Non-Member Price|
Recently, we have introduced two types of nonlinear image correlators with superior performance com to that of the conventional linear optical corre1ators.1 The first type of nonlinear optical processor is joint Fourier transform based which uses nonlinearity at the Fourier plane to threshold the joint power spectrum.4 We have shown that compared with the linear correlator, the nonlinear correlator can provide a much higher autocorrelation peak intensity, smaller correlation sidelobes, narrower correlation width, and better discrimination sensitivity. The second type of nonlinear optical correlator is the nonlinear matched ifiter based correlator13. The bandpass nonlinear matched filter is computed by expressing the linear matched ifiter as an amplitude and phase modulated bandpass waveform using the amplitude and phase of the Fourier transform of the reference function. The modulated bandpass real function is then transformed using a general type of nonlinearity.