A new preprocessing algorithm for detection of a small dim target in an IR image squence

In this paper, we propose a new algorithm to preprocess an IR image sequence for the detection of a small target which is undistinguishable because it is dim and embedded in strong clutter. This algorithm integrates the background suppression based on wavelet transform and the improved frame difference to enhance the target intensity, suppress the cluster, and improve the SNR. Experimental results with a real IR sequence in OTCBVS Benchmark Dataset Collection illustrate that this approach is very feasible and effective to improve the SNR and make the dim target well distinguishable.