基于特征融合的改进粒子滤波目标跟踪算法

针对复杂背景下单一的颜色特征不能准确跟踪目标的问题,提出了一种改进的目标跟踪算法。该算法利用跟踪目标的颜色特征和运动边缘特征来表示目标。在粒子滤波的框架下融合特征信息从而进行目标跟踪,能够有效地避免单一颜色特征在跟踪过程中受到相似背景、遮挡等问题的干扰。通过与基于单一颜色特征跟踪实验误差数据的分析,实验结果表明该算法在复杂背景以及目标遮挡等情况下能达到较好的目标跟踪效果,实现目标的准确跟踪。

Improved particle filter object tracking based on feature fusion

Under a complex environment,object tracking can’t be accurately processed through using single color feature.Thus,an improved object tracking method is proposed.The proposed method uses color feature and motion feature to present the interested target,and conducts target tracking by fusing the features in the frame of particle filter.This method can effectively avoid the various interferences in the object tracking process with single color feature,such as similar backgrounds and occlusion.Comparing the tracking errors of the proposed method with the method based on single color feature,the experiment results show that the improved method can achieve better results under the occlusion and the changes in the complex background,etc.