Infrared Small Target Detection Based on Weighted Variation Coefficient Local Contrast Measure

Infrared small target detection is one of the key technologies in IR guidance systems. In order to obtain high detection performance and low false alarm rates against intricate backgrounds with heavy clutters and noises, an infrared small target detection method based on weighted variation coefficient local contrast measure is proposed in this paper. For the raw infrared image, the variation coefficient local contrast map is calculated firstly, which can extract the local contrast features of different background regions better. Then, the modified local entropy is used as weights for the contrast map to enhance the target further. After that, a simple adaptive threshold is applied to segment the target. Experimental results on four sequences compared with seven baseline methods demonstrate that our method not only has better detection performance even if with strong clutters, but also can suppress the interference simultaneously.