Improved Kernel Correlation Tracking Algorithm Based on Invariant Moments and Kalman Prediction
Aiming at the problem that the KCF(Kernel Correlation Filter)algorithm is prone to losing target when the target is severely occluded,a Kernel Correlation Filter tracking algorithm combining the invariant moment features and Kalman prediction is proposed.The similarity metric value is calculated by using the invariant moment features of the template image and the current frame image,and then the target occlusion judgment mechanism is set according to the change of the similarity metric value under different occlusion conditions.When the target is not occluded,the KCF algorithm continues to track the target.When the target is occluded,the Kalman filter is used to predict the target position,and the KCF algorithm uses the predicted position information to continue tracking after the target recurs.Finally,the OTB-2015 dataset is used to test the effectiveness of the algorithm.The experi-mental results show that the tracking accuracy and tracking success rate of the proposed algorithm are improved by 12.95%and 13.28%,respectively,compared with the KCF algorithm,which effectively improves the target tracking performance under occlu-sion conditions.
object trackingcorrelation filterKalman filterinvariant moment feature
万方数据
维普
