基于嵌入式边缘设备的端-边-云视频流分析系统构建  

Construction of Mobile-Edge-Cloud Video Analysis System Based on Embedded Edge Devices

作　　者：朱锦辉[1] ZHU Jinhui(No.91404 Troops of PLA,Qinhuangdao 066001)

机构地区：[1]中国人民解放军91404部队,秦皇岛066001

出　　处：《舰船电子工程》2022年第7期110-115,共6页Ship Electronic Engineering

摘　　要：智慧军营是在继智慧城市之后产生的一个全新命题。近几年来,卷积神经网络的崛起推动着视频流实时分析应用快速发展,并造就了高效率、高准确度的智能分析水准。在智慧军营的建设方面,在端-边-云计算架构之上对监控摄像头的视频流进行灵活的实时智能分析逐渐变得不可或缺。论文以军营中的视频流车辆信息实时智能分析应用为例,提出了一项端-边-云计算架构之上的实时性优先的智能组件复用策略VideoEmbedded,同时针对大规模视频流实时智能分析应用场景,基于Docker容器技术和K8s容器编排技术,设计与实现了一套具有良好弹性伸缩与自动化运维能力的端-边-云系统。Smart military camp is a new proposition after smart city.In recent years,the rise of convolutional neural network has promoted the rapid development of real-time video analysis,and created the intelligent analysis level of high efficiency and high accuracy.In the construction of smart military camp,flexible real-time analysis of video streams from surveillance cameras based on Mobile-Edge-Cloud computing architecture is becoming increasingly indispensable.In this article,the application of real-time video analysis for vehicle information is taken as an example,VideoEmbedded,a real-time intelligent component reuse of priority strategy based on Mobile-Edge-Cloud computing architecture is proposed.At the same time,a Mobile-Edge-Cloud system with good ability of flexibility and automation of the operational is designed and implemented based on Docker containers and K8s con⁃tainers orchestration,for large-scale real-time video analysis application.

关 键 词：视频流分析 嵌入式边缘设备 端-边-云计算架构 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]