基于粒子滤波的双模图像融合自适应跟踪

针对视频跟踪中仅利用目标的单特征容易导致跟踪失败的问题,提出一种基于粒子滤波的可见光与红外序列图像相融合的自适应目标跟踪算法;该算法在粒子滤波跟踪算法框架下,根据单一信源运动目标序列图像的品质因子,利用自适应加权融合策略重构双模序列图像的特征选择机制,建立了基于自适应融合算法的系统观测概率模型和状态空间层次采样多特征融合跟踪算法,实现了对双模序列图像的融合以及对运动目标的稳健跟踪;跟踪试验结果表明,该算法可以有效实现对运动目标的稳健、准确跟踪。

Adaptive Tracking of the Double Modules Image Fusion Based on Particle Filter

For the problem of tracking failure easily in video target tracking using single feature,an adaptive target tracking algorithm of the visible and infrared image sequences features fusion is proposed based on the particle filter in this paper.The algorithm based on the particle filter in target tracking framework uses the adaptive weighted fusion strategy to reconstruct double modules sequence image feature selection mechanism according to a single source moving targets sequence image quality factor,and establishes the probability module of systematic observation based on adaptive fusion and multi-feature fusion tracking algorithm based on level of sampling in the state space,and realizes the double modules sequence image fusion and moving target stable tracking.The results of test show that the method mentioned above can achieve the stability and accuracy in target tracking.