摘要
工业视频网络是由工业网络系统现场层的视觉感知终端组成的网络,是实现工业网络系统泛在感知的重要基石.通过支持边缘计算层和现场设备层之间的交互和物联,工业视频网络将独立的视觉传感器单元无线连接、边缘处理,以实现空间分散下的协作监控和精确感知.它具有感知维度高,网络动态性强,感知与传输、计算、存储紧密耦合等突出特性.如何在计算、网络、存储资源受限环境下实现终端压缩提纯、边缘协作处理、云端敏捷分析,是这类系统研究的新挑战.本文首先简述工业视频网络的定义和主要特征;其次分析工业视频网络智能感知面临的挑战和关键问题;然后综述基于边缘计算的工业视频网络智能感知关键技术的研究进展;最后对工业视频网络智能感知的未来研究方向和潜在应用前景进行总结和展望.
Industrial video network is a network composed of visual perception terminals at the field layer of industrial network systems,which is an important cornerstone for achieving ubiquitous perception in industrial network systems.By supporting the interaction and connection between the edge computing layer and the field device layer,the industrial video network wirelessly connects and processes the edges of independent visual sensor units to achieve collaborative monitoring and accurate perception under spatial dispersion.It has outstanding characteristics such as high perception dimension,strong network dynamics,close coupling between perception and transmission,computing and storage.How to achieve terminal compression and purification,edge collaborative processing,and cloud agile analysis under limited computing,network and storage resources is a new challenge in the research of such systems.This paper firstly describes the definition and main characteristics of industrial video network;Secondly,the challenges and key issues of industrial video network intelligent perception are analyzed;Thirdly,the research progress of intelligent perception key technologies for edge computing based industrial video network are summarized;Finally,the future research direction and potential application prospect of industrial video network intelligent perception are summarized and prospected.
作者
涂静正
温晓婧
陈彩莲
关新平
TU Jing-Zheng;WEN Xiao-Jing;CHEN Cai-Lian;GUAN Xin-Ping(Department of Automation,Shanghai Jiao Tong University,Shanghai 200240;Key Laboratory of System Control and Information Processing,Ministry of Education of China,Shanghai 200240;Shanghai Engineering Research Center of Intelligent Control and Management,Shanghai 200240)
出处
《自动化学报》
北大核心
2025年第8期1715-1738,共24页
Acta Automatica Sinica
基金
国家自然科学基金(62025305,62432009,62227811)资助。
关键词
工业网络系统
工业视频网络
智能感知
边缘计算
多目标跟踪
Industrial network systems
industrial video network
intelligent perception
edge computing
multi-object tracking
