基于多特征融合的均值迁移粒子滤波跟踪算法

仅用单一的颜色特征进行跟踪是大多数跟踪算法鲁棒性不高的主要原因。针对此问题，该文提出一种多特征融合跟踪算法。该算法利用颜色和纹理特征表示目标，通过均值迁移和粒子滤波算法进行特征融合，有效地避免了单一颜色特征在光照变化和背景相似情况下的不稳定问题。将两种常用的融合策略结合，减轻了粒子的退化现象，提高了算法效率。实验结果表明该算法提取的目标特征具有较强的鲁棒性，能实现复杂场景下的目标跟踪。

Particle Filter and Mean Shift Tracking Method Based on Multi-feature Fusion

Object tracking by using single color feature results in a poor performance in robustness. To solve this problem, an object tracking method based on multi-features fusion is presented. The proposed method uses the color and texture features extracted by Local Binary Pattern(LBP) to present the interested target, performs a feature fusion in mean-shift and particle filter algorithms, and efficiently avoids the unstable problems via using single color feature for representation. The two common used fusion rules are used,thus overcoming the degeneracy problem and resulting in low computational cost. Experimental results indicate the proposed method is more robust to present object and has good performance in complex scene.