基于Mean-shift改进的自适应目标跟踪算法  

Improved Adaptive Target Tracking Algorithm Based on Mean-shift

作　　者：张伟 李绍铭 王勇 ZHANG Wei;LI Shao-ming;WANG Yong(School of Electrical and Information Engineering,Anhui University of Technology,Maanshan Anhui 243032,China)

机构地区：[1]安徽工业大学电气与信息工程学院,安徽马鞍山243032

出　　处：《兰州工业学院学报》2020年第6期72-77,共6页Journal of Lanzhou Institute of Technology

摘　　要：针对传统Mean-shift跟踪算法在复杂背景下存在跟踪能力不足的缺点,提出一种基于分块背景模型实时更新的背景模型,在更新过程中不断改变阈值和目标跟踪框尺寸.以Mean-shift跟踪模型为基础,建立背景模型和目标模型并引入实时更新背景模型和尺寸自适应算法,提高目标信息和背景信息的区分度以及复杂背景下对目标模型的辨识能力.结果表明:改进后的算法具有对视频序列中运动目标有效实时跟踪的能力.Aiming at the shortcomings of insufficient tracking ability of the traditional Mean-shift tracking algorithm in the complex background,a real-time updating background model is proposes based on the block background model and continuous change of the threshold and target tracking frame size during the update process.Based on the Mean-shift tracking model,the background model and target model are established,and the real-time update background model and size adaptive algorithm are introduced to improve the distinction between target information and background information and the ability to identify target models in complex backgrounds.Experiments show that the improved algorithm has the ability to effectively track moving targets in video sequences in real time.

关 键 词：MEAN-SHIFT算法 目标跟踪 背景更新 尺度自适应 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]