Infrared Small Target Detection Based on Anisotropic Contrast Filter

Local contrast filter makes use of the contrast between the target and its surrounding background for target detection. While enhancing SNR of the target, it inevitably leads to the enhancement of clutter such as cloud edges and produces false alarms. Inspired by the observation that infrared small targets shows strong contrast with the surrounding background in all directions, while false alarms caused by cloud edge usually has only strong contrast in one direction, this paper combines anisotropic filter with contrast filter and proposes an anisotropic contrast filter for infrared small target detection. Compared to the original local contrast filter, our method further considers the contrast in different directions to achieve better robustness. Experimental results show that the proposed method can effectively suppress false alarms generated by original contrast filter and achieve better performance.