Abstract: In order to reduce the computation burden and tracking deviation of target recognitions owing to moving targets with dynamic background, a fast target lock tracking method has been proposed, which is based on Hu_Fourier feature descriptors target recognition algorithm and fixed template sliding confidence least square trajectory estimation algorithm. Firstly, according to the proposed Hu_Fourier descriptors, the counter features are extracted quickly and accurately, by which all targets in the image are recognized on the basis of the weighted distance target recognition algorithm. And the centroid coordinates of the targets are calculated. Then, the target is locked exactly using the proposed fixed template sliding confidence criterion and the recognized target centroid coordinates. The target template is updated by the locked target feature. At last, using the best linear approximation method, the recognition region in the next frame image is determined. The theoretical analysis and experimental results show the robust and efficiency of algorithm proposed.

Key words:  Hu_Fourier feature descriptors, fixed template sliding confidence, moving target tracking, dynamic background

摘要： 为改善动态背景运动目标跟踪的精度和计算性能，提出一种Hu_Fourier特征描述子的目标识别算法和固定模板滑动置信度最佳线性逼近法的目标锁定跟踪算法。依据提出的Hu_Fourier特征描述子，快速准确地提取出识别区域运动目标的轮廓特征，结合加权距离目标识别算法，识别出图像中所有目标，计算目标的质心坐标。利用提出的固定模板滑动置信度准则，结合识别目标的质心，锁定目标，更新目标模板。依据最佳线性逼近法，确定下一帧图像的目标搜索区域。理论分析和实验结果验证了该目标锁定跟踪算法的鲁棒性和有效性。

关键词: Hu_Fourier特征描述子, 固定模板滑动置信度, 运动目标跟踪, 动态背景