Performance of the SMLR deconvolution algorithm

A performance analysis is presented for false alarms, correct detections, and the resolution of the suboptimal maximum-likelihood deconvolution (MLD) algorithm, called the single most likely replacement (SMLR) algorithm. It is assumed that the source wavelet and statistical parameters are given a priori. It is shown analytically that the performance improves as the signal-to-noise ratio (SNR) increases and as the mainlobe width of the normalized autocorrelation function of the source wavelet decreases. For the same performance, a higher SNR is required as the mainlobe width of the normalized autocorrelation function increases. Some simulation results which are consistent with the analytic results are presented