Abstract
Optimal recognition of patterns with respect to noise tolerance is obtained when matched filtering is applied with colored-noise compensation, for which input-noise spectral characteristics have to be known in advance. Conventional joint transform correlator systems provide such optimality only when the input noise is white. We propose general means to facilitate colored-noise tolerance by incorporating either a priori or adaptive noise compensation in joint transform correlator–based matched filtering schemes. Adaptive compensation for colored noise may be achieved in real-time operation by estimation of the noise power spectral density distribution from input power spectra information.
© 1995 Optical Society of America
Full Article | PDF ArticleMore Like This
Hanni Inbar and Emanuel Marom
Opt. Lett. 20(9) 1050-1052 (1995)
Hanni Inbar and Emanuel Marom
Opt. Lett. 18(19) 1657-1659 (1993)
P. García-Martínez, J. L. Martínez, M. M. Sánchez-López, and I. Moreno
Appl. Opt. 49(26) 4866-4873 (2010)