Desigh And Implementation Of Computer Room Monitoring System Based On Edge Computing Platform

With the continuous development of computer room monitoring technology,more and more unattended computer room monitoring systems have replaced the inefficient manual inspection work of the computer room.Ordinary computer room monitoring systems generally only have computer room environment monitoring sensors and surveillance cameras,and most of the host hardware operating status in the computer room adopts network operation and maintenance software monitoring methods.In the actual operation and maintenance process,sometimes the alarm information reported to the operation and maintenance software is incomplete due to the underlying hardware problems of the device,resulting in an error indicator on the device panel but no network alarm information is generated.This paper designs and implements a computer room monitoring system based on video recognition monitoring equipment panel indicator status.Considering that the location of the computer room is widely distributed,and the method of transmitting the recorded video to the cloud server for identification and analysis cannot solve the problems of time delay and timeliness,this paper designs a computer room monitoring system based on an edge computing platform.Among them,the edge computing end adopts a hardware platform with high energy efficiency ratio based on the RISC-V core,and designs an edge controller that integrates temperature,humidity,and smoke sensors,monitors the environmental information of the computer room,and implements local running video analysis algorithms,which effectively reduces The network load of the monitoring system reduces the delay of network transmission and improves the ability of the computer room to deal with emergencies.This article mainly completes the following tasks:(1)Design and implement an edge controller based on the Kanzhi K210 chip,which can collect data from temperature,humidity,and smoke sensors,and detect the environmental status of the computer room through data analysis.It can also capture images of the host device panel indicator lights through image processing.The algorithm recognizes the indicator light and judges its status,and finally reports the processing result to the cloud server through the 4G module.(2)This article optimizes the indicator status recognition algorithm for low-resolution cameras.Based on cost control,the system uses a low-cost camera,the image quality is poor,and there is a serious color offset,this paper proposes a targeted indicator status recognition program,and the experimental results show that the program can better indicate the status of the light is recognized.(3)Designed and implemented a computer room monitoring and management system based on a cloud platform.The system mainly completes two tasks.First,it receives and stores data in the computer room,analyzes and counts the operation of the computer room,and issues instructions to the computer room.Secondly,for images that are difficult to recognize by the edge controller due to factors such as illumination and shooting angles,the cloud the platform uses its computing advantages to further process it.