Infrared Dim Small Target Tracking Based on Inter-frame Consistency Under Complex Background

Based on the inter-frame correlation and continuity of target features, a method of dim small target tracking in the infrared sequence is proposed. To depress the noise, the spatial hard constraint is used to eliminate the noise that is far away from the target, and the credibility soft constraint evaluates the indistinguishable noise that is close and similar to the target. Aiming at the disappearance of the target caused by the over-bright region, the prior information of the target trajectory of known frames in the past is fully excavated to improve the continuity of target tracking. All the above contributions have been integrated into a unified framework, which is a probabilistic pipeline filtering algorithm with trajectory prediction. The experimental results on five different infrared image sequences prove that the proposed algorithm can effectively improve the robustness and accuracy of tracking in complex environments.