摘要
针对基于模板的目标跟踪算法存在模板冗余高、难以适应非刚性目标外观多变的问题,提出一种基于粒子滤波的模板在线聚类目标跟踪方法.首先建立用于描述目标和背景的正、负模板集,然后抽取候选粒子,使用候选粒子与正、负模板集的类内距离以及正、负模板集之间的类间距离来构建似然函数,最后依据最大后验概率准则确定最佳候选粒子作为跟踪结果.根据视频序列中连续变化的目标状态,将一定范围内的相似目标状态视为一个状态类,确定当前状态类的聚类半径.采用均值漂移算法对正模板集及最近几帧跟踪结果进行聚类,并将聚类后的中心集作为新的正模板集.实验表明,该算法能保留目标不同的外观状态,在复杂情况下仍能准确跟踪目标.
The template of most tracking algorithms is difficult to suit appearance changes of non-rigid objects due to high redundancy. To deal with the problem, a tracking method with online templates clustering based on particle filter is proposed. Positive and negative template sets are established to describe the target and background. Candidates are then extracted based on the dynamic model. A likelihood function is built based on the within- class distance between the candidate and template sets, and the between-class distance between positive sets and negative sets. The best candidate is considered the tracking result according to the maximum a posteriori probability （MAP）. The main idea of online template clustering is as follows： First, the cluster radius is determined by the state class produced from a series of continuous target states in a certain range. Second, the positive template set combined with the recent several tracking results are used for clustering by using the mean shift iterative method and the above cluster radius. Third, the updated positive template set consists of the new cluster centers. In complicated situations, experiments show that this method can retain different appearance states of target and track the target accurately.
出处
《应用科学学报》
CAS
CSCD
北大核心
2016年第4期430-440,共11页
Journal of Applied Sciences
基金
国家自然科学基金(No.61203273)
江苏省自然科学基金(No.BK20141004)
江苏省普通高校自然科学研究基金(No.11KJB510009
No.14KJB510019)
江苏省"信息与通信工程"优势学科项目
江苏省大学生实践创新训练计划项目基金(No.201510300036Z)资助
关键词
粒子滤波
模板
类间距离
类内距离
均值漂移
在线聚类
particle filter, template, between-class distance, within-class distance, mean shift, online clustering
