一种鲁棒的自适应更新策略的弹载计算机红外目标跟踪算法

Robust Adaptive Updating Strategy for Missile-borne Infrared Object-tracking Algorithm

针对目前弹载红外目标跟踪算法对快速机动目标适应性较差的问题,本文在现有模板跟踪算法的基础上提出了一种鲁棒普适的自适应更新策略。该策略利用多个不同学习因子的更新模型,通过对不同模型下的跟踪置信度进行分析,自适应地对模型进行更新,有效地解决跟踪过程中目标剧烈外观变化或微弱位移变化。大量的定性定量实验表明本文所提的算法的跟踪性能超过现有的大多数算法,在现有弹载红外目标跟踪板硬件余量有限的情况下也能实现稳定跟踪复杂背景下的目标。同时,本文算法已经在挂飞测试中获得较理想的效果,适合军事装备的应用。

To mitigate the poor adaptability of the missile-borne infrared object-tracking algorithm for a faster maneuvering,a robust universal adaptive updating strategy based on the existing template tracking algorithm is proposed.The strategy model with different learning factors is first used,the analysis of the tracking confidence are analyzed under different models,and then the adaptive model can be updated to solve the severe appearance changes and weak displacement in the process of object tracking.The qualitative and quantitative experiments show that the tracking performance of the proposed algorithm is better than most existing algorithms,which can also achieve a stable object under a complex background in the existing missile-borne hardware margin.In addition,our algorithm has also achieved the ideal results in the captive flight test,which is suitable for military equipment applications.