基于MeanShift的目标平移与旋转跟踪

针对目前多数实时跟踪算法只能跟踪目标平移运动,不能跟踪旋转运动的问题,提出一种基于Mean Shift的快速旋转跟踪算法。该算法以目标区域的梯度方向分布(直方图)为特征,构造了可用Mean Shift算法寻优的相似度函数,将旋转跟踪转化为寻优问题,并利用Mean Shift寻优过程收敛速度快的特点,有效跟踪目标旋转。又提出交替迭代的方法,将旋转跟踪与Meer的平移跟踪算法融合起来,构造了可以同时跟踪目标旋转和平移完整跟踪算法。

Mean Shift Based Targets' Rotation and Translation Tracking

The present real-time tracking algorithm can perform well in term of translation tracking but can seldom do so in rotation tracking. A rotation tracking algorithm was proposed, which utilized the gray gradient direction distribution of the target region as the feature and constructed a similarity function that can be optimized by Mean Shift method. Therefore the rotation tracking was transformed into an optimization problem. Due to the fast convergence of the Mean Shift, this algorithm can be run in real-time. Combining the rotation tracking with the translation tracking algorithm proposed by Meer, a whole algorithm was obtained by alternate iteration, which can track both translation and rotation of the targets.