摘要
针对运动目标跟踪的过程中,运动轨迹发生改变或运动过程出现目标遮挡时,传统的跟踪算法容易出现跟踪丢失或跟踪效果较差等情况,文章提出了一种基于交互多模型Kalman滤波算法和Meanshift滤波算法的融合方法,对运动目标进行位置预测估计,解决目标遮挡问题,提高目标跟踪精度。实验结果证明,该改进算法能够有效地进行目标定位跟踪,并且具有快速性。
Aiming at the situation that the moving trajectory changes or the target occludes during the tracking of the moving target,the traditional tracking algorithm is prone to track loss or poor tracking effect.A fusion method based on the interactive multi-model Kalman filtering algorithm and the Meanshift filtering algorithm was proposed,which could estimate the position of moving targets,solve the problem of target occlusion,and improve the accuracy of target tracking.Experimental results showed that the improved algorithm couldeffectively perform target positioning and tracking,and itwasfast.
作者
沈志伟
赵晨希
侯君怡
余雷
SHEN Zhiwei;ZHAO Chenxi;HOU Junyi;YU Lei(School of Mechanical and Electrical Engineering,Soochow University,Suzhou 215021,China)
出处
《电工技术》
2020年第18期41-43,共3页
Electric Engineering
基金
苏州大学2019年国家级大学生创新创业训练计划立项项目(项目号:201910285002X)
苏州大学一流本科专业计划项目
江苏省研究生科研与实践创新计划项目(项目号:KYCX19_1927)。
