Detecting Infrared Maritime Dark Targets Overwhelmed in Sunlight Interference by Dissimilarity and Saliency Measure

Efficient infrared (IR) maritime target detection in sunlight interference conditions (SIC) is a severe challenge task in maritime distress target search and rescue systems. To solve this problem, the main contribution of the paper effectively proposed a dissimilarity and saliency measure method. Firstly, based on the different orientations feature between the target and the background, the suspected target patches are extracted by revised visual attention model (RVAM). Secondly, considering the nonlocal characteristics of the target, a novel nonlocal dissimilarity measure is proposed to achieve clutter suppression and target enhancement. Finally, since the intensity changes of target and background in continuous frames, the multi-frame saliency measure method is applied to remove clutter noise with low intensity similar in the target shape. Experimental results demonstrate that the proposed method is superior to the state-of-the-art methods concerning detection validity and robustness for IR maritime target detection in SIC.