基于异步时域事件的高速运动目标的跟踪算法  

High-speed moving target tracking algorithm based on asynchronous transient event

作　　者：杨帅[1] 薛永江[1] 王瑞昆[1] 

机构地区：[1]天津工业大学计算机科学与软件学院,天津300387

出　　处：《计算机时代》2016年第6期72-74,79,共4页Computer Era

基　　金：天津市科技重大专项与工程项目(15ZXHLGX00300)

摘　　要：异步时域事件是由异步时域视觉传感器获得,其数据具有低冗余特点,适宜于高速运动目标的跟踪。基于异步时域事件的跟踪算法从数据源上降低了数据冗余,克服了基于"帧"数据的传统跟踪算法需要处理大量无用信息的缺点,极大地提高了跟踪速度。文章提出的算法充分考虑了异步时域事件的数据集中性,在目标跟踪中使用按周期更新的策略,降低了算法的时间复杂度,提高了跟踪速度,并且跟踪轨迹更加平滑。Asynchronous transient event is obtained by asynchronous time-domain visual sensor, which has a low data redundancy feature, and is suitable for tracking high-speed moving targets. Tracking algorithm based on asynchronous transient event reduces data redundancy from the data source, and overcomes the shortcomings of traditional tracking algorithm based on the ＂frame＂ data which has large volume of useless information, and greatly improves the tracking speed. The proposed algorithm fully considers the concentration of asynchronous transient event, so use the periodically updated in target tracking, and greatly reduce the time complexity of the algorithm, so that the tracking speed is improved and the tracking trajectory is smoother.

关 键 词：异步时域事件 目标跟踪 高速 时间复杂度 更新策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]