Performance comparison of a linear parametric noise estimation Wiener filter and non-linear joint transform correlator for realistic clutter backgrounds

摘要

It has been shown previously that a linear Wiener filter is capable of detecting a target in severe clutter backgrounds by utilising a parametric model of the clutter power spectrum in its filter transfer function. In this paper the performance of the linear Wiener filter is compared to that implemented in a non-linear joint transform correlator in which the entire current input scene is used as an approximation for the clutter background. Realistic clutter backgrounds are employed in the tests that cover a range of natural scenery likely to be encountered in practice. The linear Wiener filter, employing a parametric model of the averaged background scenes, is shown to outperform the non-linear filter in most cases. Brief consideration is also given to the relative merits of implementation of these two filters in both optical and digital correlators.