基于选择性模型更新与卡尔曼滤波的目标模型更新算法

选择性模型更新算法不能准确地更新目标模型,在外观变化、遮挡、场景光线变化等因素影响的运动目标跟踪中,不能有效地处理目标模型。因此,提出了一种选择性模型更新与卡尔曼滤波的目标模型更新算法。根据可靠性阈值和分量更新比例精确选取更新分量,并与Kalman滤波相结合,对目标模型分量进行预测,根据不同干扰和目标外形变化,将两种算法的跟踪结果线性加权得到新的跟踪目标模型。实验结果表明该算法具有良好的跟踪效果。

A Target Model Update Algorithm Based on Selective Model Updating and Kalman Filtering

S elective model update algorithm can not accurately update moving target model, cannot effectively process the target model, in the scene of target appearance changes, occlusion, lighting changes and other factors. Therefore a target model update algorithm based on selective model updating and kalman filtering is proposed. It is accurately select the update component, according to reliability threshold and the proportion of component updates, and combination the Kalman filter, predict the component of target model. The new tracking target model is get by two algorithms tracking the results of linear weighted, depending on the external interference and the target shape changes. Experimental results show the better stability and robustness of the proposed algorithm.