摘要
跨摄像机行人再识别是智能视频监控系统中一个基本的任务,跨摄像机行人再识别就是鉴别出行人在摄像机网络中不同摄像机里出现的确认的过程,近些年其已成为计算机视觉领域的研究热点并受到了许多研究领域的关注。摄像机间拍摄距离、拍摄角度、分辨率等差异和时空信息的不确定性使得跨摄像机行人再识别面临巨大的挑战,同时也激励着研究者不断发现新算法来解决问题。此文通过对近些年公开发表的论文进行梳理,阐述了跨摄像机再识别目前面临的困难、当前的研究现状和将来的研究趋势。
The cross-camera person re-identification is a basic function in the smart video monitoring system.Cross-camera person re-identification is the process of identifying the presence of persons in different cameras in the camera network.In recent years,it has become a research hotspot in the field of computer vision and received attention in many research areas.The differences in shooting distances,angles,resolutions between cameras and the uncertainty in spatio-temporal information bring enormous challenges to cross-camera person re-identification,but it also motivates researchers to constantly find new algorithms to solve the problem.This paper sorts out the papers published in recent years,and illustrates the current difficulties in cross-camera re-identification,current research status and future research trends.
作者
谭飞刚
廖全蜜
李健艺
Tan Feigang;Liao Quanmi;Li Jianyi(School of Traffic and Environment,Shenzhen Institute of Information Technology,Shenzhen,Guangdong,China 518172)
出处
《深圳信息职业技术学院学报》
2018年第5期72-78,共7页
Journal of Shenzhen Institute of Information Technology
基金
深圳信息职业技术学院校级青年培育项目(QN201709)
关键词
行人再识别
跨摄像机
无视野重叠
智能视频监控
目标再识别
person re-identification
cross-camera
non-overlapping view
intelligent video monitoring
object reidentification
