Abstract: To deal with the lower detection probability in the low SNR by the traditional pipeline filter algorithm resulting from marginal noise interference in complicated IR background, a variable weighted pipeline filter algorithm is presented for detecting small targets in IR image sequences. The adaptive learning scheme is employed to modify the pipeline center position, i.e., variable weighted pipeline center coordinates, according to targets’ positions. This method can restrain effectively the interference caused by marginal noise. Compared with the traditional detection algorithm, this method is more suitable for dealing with failure detection and targets disappearing resulting from marginal pipeline noise. Experiments on long-rang IR image sequences show that the proposed algorithm can detect small targets within the IR image sequence efficiently with the SNR higher than 1.5.

Key words: Robinson Guard filter, IR image sequence, small targets detection, pipeline filter