Intelligent decision-making technology based on infrared image complexity

With the rapid development of infrared-guided missile detection and anti-jamming technologies, traditional evasion and defense measures are increasingly ineffective in countering complex air combat environments, severely restricting the survival capabilities of combat aircraft. In response to the issue of autonomous evasion of infrared-guided threats, this paper proposes an intelligent jamming strategy generation framework based on infrared image complexity. First, an effective feature space for infrared countermeasure scenarios is constructed, including basic statistical features, gradient features, and texture features. Based on this, the local background occlusion degree, the global background confusion degree, and the target information loss degree are proposed to quantify the effectiveness of jamming strategies in different situations. Second, a KAN network is introduced to establish the mapping relationship between engagement scenarios, optimal jamming strategies, and complexity, thus overcoming the limitations of traditional empirical decision-making. The simulation results show that this method achieves a success rate in jamming 78%, effectively evaluates the jamming effects in different situations, and selects the optimal jamming strategy. The method improves the intelligence level and decision-making efficiency of infrared jamming, and compared to other infrared jamming decision-making methods, it demonstrates better performance in model accuracy, interpretability, and generalizability.