基于多特征自适应融合的分类采样跟踪算法

针对目标跟踪中的场景易变和目标模板不稳定等问 题,提出了一种基于多特征自适应 融合的分类采样跟踪算法。算法利用密集特征信息将目标模板用多个重叠子区域划分,每 个子区域对应一个多特征采样窗口。利用多特征自适应融合构造强可区分性的目标模型,最 大程度地提高各子区域之间的互补性,以增强目标模板的区分能力。在粒子滤波(PF)框架下 , 多特征自适应融合策略提高了目标观测质量,保证跟踪的持续稳定。实验结果表明,本文所 提算法具 有良好的目标跟踪性能,并对动态场景、目标形变及遮挡情况具有较好的跟踪准确性和鲁棒 性。

Classified sampling tracking algorithm based on adaptive multiple features fusi on

Aimed at the varied envir onment and unstable object template in visual tracking,a classified sampling tra cking algorithm based on adaptive multiple features fusion is presented in this paper.The proposed method uses some overlapping sub-regions to divide the target model by the information of dense features,and each sub-region correspo nds to a multi-feature sampling window.The discriminative object template is construct ed by the adaptive multiple features fusion method,and this proposed strategy maximally improve the co mplementary of sub-region and the discriminability of template.The fusion strategy can improve the quality of measurement,and ensure sustainable and stable tracking in a particle filter fra mework. Experimental results show that the proposed method has good performance,and is more robust and stable to pose variation,dynamical background and occlusion than other methods.