Infrared Small Target Detection Through Multiple Feature Analysis Based on Visual Saliency

Infrared small target detection in extreme environments such as low illumination or complex background with low signal clutter ratio is of crucial significance and counted as a difficult task in infrared search and tracking systems. In this paper, an effective infrared small target detection method is proposed based on the human visual system characteristics and multiple feature analysis. By using the contrast mechanism and visual attention mechanism, spatial gray level-based feature map and saliency extraction based feature map from frequency domain are obtained. Based on the characteristics of infrared dim target images and human visual attention mechanism, the target saliency features are revealed through the feature analysis in the spatial domain and frequency domain respectively. The saliency features from each feature map are applied to the final saliency map. By this means, the background clutter and noise are inhibited and the targets are distinct for the various scenes in the infrared images. The experimental results show that the proposed method has a robust and effective performance in terms of detection and false alarm rates. Comparing with the other methods in the experiments, the proposed method is feasible and adaptable in the various scenes of infrared images.