一种复杂背景下非线性图像目标跟踪算法

A nonlinear image target tracking algorithm in complex background

针对复杂背景下的图像目标准确跟踪问题,提出一种基于颜色、SURF和ORB特征的粒子跟踪算法,该算法利用多特征融合描述目标,克服复杂背景带来的影响,同时,提出一种自适应模板更新策略,解决了复杂背景下的模板衰退问题。实验表明,该算法在存在背景干扰、目标旋转、遮挡等复杂场景下具有较强的准确性。

Aiming at the problem of accurate image target tracking in complex background, this paper proposes a particle tracking algorithm based on color, surf and orb features. The algorithm uses multi feature fusion to describe the target to overcome the influence of complex background. At the same time, an adaptive template updating strategy is proposed to solve the problem of template fading in complex background. Experimental results show that the algorithm has a strong accuracy in complex scenes such as background interference, target rotation, occlusion and so on.