An improved method for infrared small target detection under various complex backgrounds

Influenced by the optics point spread function of the thermal imaging system at a remote distance, the small target is analogous to isotropic Gaussian-like shape, whereas background clutters are usually the shape of the local orientation clutter. To tackle this problem, an improved infrared target detection method is proposed in this paper, which adopts multiple channels in the process of image enhancement based on the characteristics of the infrared image to make it adaptive to various types of complex background clutters and to improve the robustness of proposed method for small target enhancement. First, the second order directional derivative filter based on the facet model is utilized to calculate the second order directional derivative filter maps from different channels respectively. Then, we compute the saliency maps for the different directions. Finally, the method fusions the saliency maps from different channels using improved image fusion algorithm. The proposed method is suitable for all kinds of complex backgrounds as prior knowledge and sensitive parameters are not required. Experimental results have demonstrated that the proposed algorithm is superior to the Saliency-Based method for the small target detection of various complex backgrounds.