False-alarm regulation for target detection in hyperspectral imaging

Classical target detection schemes are usually obtained deriving the likelihood ratio under Gaussian hypothesis and replacing the unknown background parameters by their estimates. In this paper, the adaptive versions of the classical Matched Filter and the Normalized Matched Filter are analyzed for the case when the mean vector of the background is unknown and has to be estimated jointly with the covariance matrix, as it is the case in hyperspectral imaging. More precisely, theoretical closed form expressions for false-alarm regulation are derived and these results are extended to non-Gaussian cases using robust estimation procedures. Finally, simulations validate the theoretical contribution.