A multiple objects tracking algorithm is introduced based on object activity classification.
针对于固定单摄像机的多人跟踪,提出一种基于活动状态分类的多目标跟踪算法,利用卡尔曼滤波对遮挡进行预测,同时结合区域匹配信息对目标的活动状态进行分类。
It introduces a new modeling method, which uses Multi-context Map(MCM) and Collaborative Linkage Map(CLM) for collaboration task.
协同工作关联图(MCM)和活动状态联接图(CLM)是协同工作过程建模新方法。
A multiple object tracking algorithm is introduced based on the object activity prediction and classification for processing this difficulty.
针对此问题,本文提出一种基于活动状态预测与分类的多目标跟踪算法,并通过对数字视频中多人的跟踪对算法进行测试。