首页 | 本学科首页   官方微博 | 高级检索  
     检索      

多运动目标探测标记及跟踪
引用本文:李金,雷燕,胡文广.多运动目标探测标记及跟踪[J].应用科技,2009,36(6):58-62.
作者姓名:李金  雷燕  胡文广
作者单位:哈尔滨工程大学,自动化学院,黑龙江,哈尔滨,150001
摘    要:针对跟踪系统对多目标跟踪以及对实时性的要求,给出了一种基于中心点和面积特征匹配的多运动目标探测标记及跟踪方法.该方法利用对多运动目标检测后的二值图像进行了连通成分标记,提出了一种新的探测搜索标记法,赋予不同连通区域不同的数字来区分,通过四连通区域法来实现.由运动目标的4个顶点来确定中心点,通过面积及中心点距离从而进一步去匹配,最后根据标记结果在原图像中准确地框定了各运动目标,从而实现对运动目标的跟踪.采用上述算法,对车辆视频进行了跟踪,取得了较好的实验结果,跟踪实验结果验证了该方法具有很好的实时性.

关 键 词:多运动目标  跟踪  序列图像  连通区域标记

Detection labeling and tracking of multi-moving targets
LI Jin,LEI Yan,HU Wen-guang.Detection labeling and tracking of multi-moving targets[J].Applied Science and Technology,2009,36(6):58-62.
Authors:LI Jin  LEI Yan  HU Wen-guang
Institution:(College of Automation, Harbin Engineering University, Harbin 150001, China)
Abstract:A real-time multi-moving target detection labeling and tracking method of moving object based on the center point and areag characteristic matching was given in the light of multi-target tracking and real-time requirements of the tracking system. After detection of the multi-moving targets, the consecutive area of the binary image is labeled. A new detection labeling method is presented by labeling the four consecutive binary images with numbers, which are acquired by the detection of multi-moving targets. The center point can be obtained by four vertexes of the moving target, and further matching can be done by measuring the areas and central distance of the moving targets. Finally, the moving targets can be traced by the labeling results of those framed moving targets. The video of vehicles was processed by the algorithm. The experimental results show that the method is effective.
Keywords:multi-moving targets  track  seqnential images  connected components labeling
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号