一种新型非线性滤波的多特征融合跟踪算法

Multi-feature fusion tracking based on new nonlinear filtering

为有效解决非线性系统的状态估计问题,提出一种新型非线性滤波算法。该算法通过在积分卡尔曼滤波中引入修正因子,对积分点进行优化重组,并采用修正后的积分卡尔曼滤波产生优选建议分布函数,较好地克服了粒子退化现象。在新算法的框架内,利用颜色和运动边缘特征作为观测模型进行视频目标跟踪,并通过D-S证据理论的方法进行权值融合,较好地克服了单一颜色特征在姿态改变、相似背景遮挡等情况下跟踪稳定性较差的问题。实验表明本方法对复杂条件下的目标跟踪问题在保持较强鲁棒性的同时,跟踪精度提升了近32%。

This paper proposed a new kind of nonlinear filtering for the state estimation of nonlinear systems.The proposed algorithm based on quadrature Kalman filter by using integral pruning factor,which optimized and reorganized the integration point.New algorithm overcame the particle degeneration phenomenon well.In the improving particle filter framework,this algorithm used color and motion edge character as observation model,and fused feature weights through the D-S evidence theory.The proposed method effectively avoided bad robust questions rosed by the single color feature in the posture change and similar feature occlusion.Experiment results indicate that the proposed method is more robust to track object in complex scene and the tracking precision ascends nearly 32%.