An Adaptive Appearance Model Update Method for Target Tracking Based on Prior Confidence

For the challenges posed by rapid changes in target scale and appearance due to the high-speed motion of aerial vehicles during the infrared imaging terminal guidance phase, an adaptive appearance model update method for target tracking based on prior confidence is proposed. This method uses OSTrack [1] as a baseline and comprehensively analyzes historical confidence variation information within a specified time window to assess the adaptability of the current appearance model and determine the timing for appearance updates. By analyzing the confidence performance of historical prediction results, reliable samples are selected to construct template information adaptively updating the target appearance model. This enhances the robustness and stability of the algorithm throughout the terminal guidance phase. Simulation results for infrared terminal guidance indicate that, compared to the baseline algorithm, the proposed method improves the success rate by 5% and increases accuracy by 4.8%. This approach provides an effective solution for enhancing the performance of infrared guidance systems.