一种基于优化模板匹配的红外目标跟踪算法

为了在红外环境下进一步提高目标图像在环境变化或受干扰时跟踪的准确性,提出了一种有效的红外目标跟踪方法.该方法基于归一化互相关距离的相似性度量方法,设计了匹配跟踪置信度,并在此基础上实现了基于匹配跟踪置信度加权的自适应模板更新算法.在模板更新方面,提出了自适应阈值调整策略,在目标跟踪过程中根据匹配置信度动态调整阈值,避免了固定阈值对跟踪精度的影响.通过实验证明,算法对环境变化的适应能力和跟踪稳定性方面得到了较大的提高.

An Infrared Target Tracking Algorithm Based on Optimal Template Matching

An effective infrared target tracking algorithm for video images is proposed to improve the accuracy of target tracking if the surroundings change or are being disturbed, based on a new similarity measurement algorithm in accordance to the normalized and cross-correlated distances with a matching confidence designed for tracking. Thus, the proposed algorithm is actually a weighted adaptive template updating algorithm based on tracking confidence. With respect to the template updating, an adaptive threshold adjustment strategy is proposed in the process of target tracking, by which the thresholds can be adjusted dynamically according to matching confidence, getting rid of the effect of fixed thresholds on tracking accuracy. Testing results showed that the proposed algorithm improves the adaptability to the change in surroundings and tracking stability greatly.