杂乱环境中改进的目标检测跟踪算法  

Target detection and tracking algorithm in cluttered environment

作　　者：段学虎 刘海英[1] 陈好男 娄海同 纪鹏 DUAN Xue-hu;LIU Hai-ying;CHEN Hao-nan;LOU Hai-tong;JI Peng(School of Electrical Engineering and Automation,Qilu University of Technology(Shandong Academy of Sciences),Jinan 250353,China)

机构地区：[1]齐鲁工业大学(山东省科学院)电气工程与自动化学院,山东济南250353

出　　处：《齐鲁工业大学学报》2022年第2期46-52,共7页Journal of Qilu University of Technology

基　　金：齐鲁工业大学国际合作项目(QLUTGJHZ2018019);国家自然科学基金青年项目(61903207)。

摘　　要：目标检测跟踪技术在当今社会的发展与应用越来越广泛。在杂乱环境中的目标检测跟踪,由于受到环境背景的影响,跟踪目标容易发生目标跟踪漂移现象、跟踪丢失或者无跟踪问题。基于经典MOSSE算法,提出改进的检测跟踪算法,实现滤波器产生误差时及时修正,很大程度上提高了检测与跟踪的准确性和实时性,降低了漂移问题导致的跟踪失败的概率。Target detection and tracking technology has become more widely used in the development and application of today's society.Target tracking is prone to drift,loss or no tracking due to the influence of background in clutter environment.Based on the classical MOSSE algorithm,this paper proposed a novel detection and tracking algorithm to achieve timely modification once the filter produced errors,which greatly improved the accuracy and real-time performance for detection and tracking,and greatly reduces the probability of tracking failure caused by drift problems.

关 键 词：杂乱环境 目标检测跟踪 MOSSE算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]