Noise-tolerant filtering schemes with joint transform correlators

Conventional joint transform correlator (JTC) configurations are optimal with respect to noise tolerance only for input scenes embedded in white noise. Different processing should be devised in the presence of colored noise, taking into account the spectral characteristics of the noise. We hereby propose general, simple and powerful means, extending the capabilities of conventional JTC systems to enable Wiener filtering (WF) as well as parametric Wiener filtering (PWF) for patterns corrupted by colored noise. Noise is explicitly accommodated for by performing mathematical operations in between the two cycles of the JTC operation. The suggested JTC-based hybrid optoelectronic implementations do not necessarily rely on a-priori knowledge of the noise power spectral density, but rather determine an estimate of this distribution in an adaptive manner, based on measurable power spectra information. Such implementations provide great flexibility and real-time adaptivity to colored noise, resulting in improved correlation performance. Mathematical analysis and computer simulations are presented, whereby noise processes with different spectral characteristics and intensities are considered.