针对模板匹配算法计算量大,难以解决目标形变、光照变化、遮挡等情况下的跟踪问题,提出了一种结合Kalman滤波的模板匹配跟踪算法。通过Kalman滤波预测下一帧图像中目标的位置,缩小目标搜索范围,以减少模板匹配计算量,满足目标跟踪的实时要求。采取自适应模板更新策略,根据目标的变化情况更新模板。在目标被短暂遮挡时,采用Kalman滤波预测目标的状态,继续稳定跟踪。实验结果表明,算法明显优于单独使用模板匹配算法,当目标被遮挡时仍可以稳定、实时跟踪,具有良好的鲁棒性。

A correlation-based tracking algorithm based on Kalman filter and adaptive template matching is discussed in this paper. The target state in the next frame can be predicted by using Kalman filter which can reduce the search area of the target detection and meet the target tracking in real time. Adaptive template update strategy which can adjust the reference template based on the changes of the tracking target is used to improve the stability of target tracking. The experimental results indicate that the algorithm is much better than the template matching method obviously. The target also can be tracked efficaciously,when it is occluded.