结合Kalman滤波器的SIFT目标跟踪算法

SIFT Tracking Algorithm Combined with Kalman Filter

针对目标图像跟踪过程中提取待匹配图像较大的特征向量时,很难满足准确性和快速性要求,提出了结合卡尔曼滤波的SIFT目标跟踪算法。算法利用Kalman滤波器对动态目标在下一帧图像中可能出现的位置,在自适应窗口中识别动态目标。实验证明,该算法可以缩短了待匹配图像的SIFT特征点提取时间,提高了目标跟踪的效率。

Target image extraction and matching image tracking larger feature vectors,It is difficult to meet the requirement of accuracy and rapidity, when extracting the eigenvectors to be matched during the target image tracking.A SIFT target tracking algorithm based on Kalman filter is proposed.The algorithm makes use of the Kalman filter to identify the dynamic target in the next frame image,which can be used to identify the dynamic object in the adaptive window.Experiments show that the proposed algorithm can shorten the time of SIFT feature points extraction and improve the efficiency of target tracking.