Infrared Moving Target Detection Method Based on Cross-Entropy-Driven Transition Zone Extraction

In the field of moving target detection and recognition for infrared image sequences, we have developed a novel approach to address infrared target detection challenges in complex dynamic scenarios. Our method leverages cross-entropy principles combined with transition zone extraction techniques to significantly improve detection accuracy and reliability. The implementation involves calculating pixel-wise entropy differences between consecutive frames to identify transition regions, followed by adaptive thresholding and noise suppression algorithms for optimal target isolation. Through extensive experimentation and validation, we have demonstrated that this approach achieves satisfactory detection results across diverse environmental conditions. Furthermore, the method exhibits excellent scalability and adaptability, enabling effective application to various infrared image sequences for enhanced detection efficiency and precision. Key algorithmic components include entropy-based frame differentiation, morphological filtering for boundary refinement, and cluster analysis for false positive reduction. We are confident in the innovation and practical value of this methodology and believe it will contribute positively to advancements in infrared target detection technologies.