Traditional　monitoring　systems　are　mostly　monocular　cameras,　the　main　function　of　which　is　to　play　and　store　video　in　the　monitoring　area　and　to　achieve　the　collection　and　judgment　of　the　information　in　the　video　through　artificial　observation.　Usually,　emergencies　that　are　not　found　and　processed　in　time　will　cause　serious　situations.　Compared　with　the　monocular　camera,　the　panoramic　camera　has　a　wide　field　of　view　and　no　blind　spots.　This　system　uses　a　panoramic　camera　to　achieve　360-degree　video　acquisition,　and　uses　deep　learning　to　process　the　video　to　obtain　the　posture　of　the　human　body,　and　then　processes　it　through　the　human　behavior　recognition　network　to　identify　the　abnormal　human　behavior　in　real　time.　Through　the　dual　warnings　of　local　and　remote　clients　and　the　artificial　re-judgment　of　the　received　abnormal　behavior　video,　the　manager　can　timely　find　and　handle　abnormal　behavior.　This　system　not　only　saves　labor　costs,　but　also　improves　the　utilization　efficiency　of　video.　The　360-degree　all-round　perspective　provides　the　area　manager　with　an　immersive　environment　reproduction,　allowing　him　to　grasp　the　details　of　the　scene　in　real　time.　This　system　provides　virtual　on-site　conditions　for　decision-making　and　handling　of　emergencies,　enabling　early　warning　and　minimizing　losses.　©　2020　Technical　Committee　on　Control　Theory,　Chinese　Association　of　Automation.