利用子区域特征进行自适应目标跟踪

Adaptive object tracking based on the sub-regions features

提出了一种基于目标区域分割的自适应运动目标跟踪算法。该算法通过K-均值聚类,将目标分割为多个子区域,根据子区域颜色特征及其分布提出了一种新的目标模型,并给出模型相似性测度准则,从而将目标模型更新问题简化为区域特征的更新,提高了模型的稳定性。同时在跟踪过程中,利用相似性测度检测目标遮挡程度,根据遮挡程度自适应地调整卡尔曼滤波器的参数和模型更新过程,提高了在遮挡情况下算法的鲁棒性。分析和实验表明,新算法能够在真实场景中准确、实时地跟踪目标,是一种有效的视频目标跟踪算法。

A visual object tracking algorithm is proposed using object region segmentation. The object model is built upon the color features and distribution of object sub-regions which is obtained using K means clustering. The up- dating of object model is simplified as the updating of object sub-regions. A similarity measure is introduced to evaluate similarity between the reference model and the candidate model. In obiect tracking, the parameters of the Kalman filter are adjusted and the object model is updated according to the similarity measure which indicates whether or not the ob- ject is occluded. The results and analysis show that the proposed method has the ability to tracking the moving obiect real time under real complex situations such as occlusion by other ones and appearance change of the moving object. The proposed method is an effective visual object tracing method.