The Design And Implementation Of Video Monitoring Alarm System For Optical Cable Line

Over the past few years,due to the advances in Internet and the construction of 5G networks,the number of optical cable lines managed by telecom operators has continued to increase,and the importance of ensuring smooth optical cable lines has continued to improve.However,in the current maintenance process of optical cable lines,the staff can only discover hidden dangers during line inspections,which not only consumes a lot of human resources,but also cannot deal with abnormal situations that occur in a timely manner.Nowadays,video surveillance has been widely used in various fields,and the development of intelligent video analysis has made video surveillance more intelligent.If intelligent video surveillance is applied to optical cable lines,manual maintenance can be assisted by technical means,which can effectively improve the efficiency of optical cable line maintenance.The topic of this thesis comes from the cooperation project of telecom operators.It designs and implements the optical cable line video monitoring and alarm system.It collects and intelligently analyzes the optical cable line monitoring video in real time.When a hidden danger in the optical cable line is found,an alarm will be generated in time,and the staff will be notified to deal with it as soon as possible.The specific work content is as follows:This thesis has analyzed and designed the optical cable line video surveillance alarm system,including system requirements analysis,system overall architecture design,and system layer design.In order to ensure the scalability of the system,according to the hierarchical structure,the system is divided into data collection layer,data storage layer,intelligent analysis layer,system service layer,system management layer,and system visualization layer.In order to prevent the optical cable from being affected by external forces and human factors,the intelligent analysis layer of the system mainly detects engineering vehicles and pedestrians.Aiming at the problems in practical application scenarios,this thesis improves and optimizes the YOLOv3-tiny detection algorithm,and proposes a deformable convolution based YOLOv3-monitor,a multi-scale detection algorithm with attention mechanism,experiments have verified the effectiveness of the optimization algorithm,which can keep the detection speed unchanged and improve the detection accuracy,and provides algorithm support for the system.This thesis has completed the realization and testing of the optical cable line video surveillance alarm system.This thesis uses target detection,microservices,streaming media and visualization technologies to realize the functions of each layer of the system.The system can collect real-time video streams and offline files,analyze the data by calling the algorithm service interface,and provide real-time alarms and GIS map display.And other services,support the management of users,equipment,and data,and realize the graphical display and streaming media playback of alarm data.Finally,the system functions and performance were tested separately to ensure the effectiveness and reliability of the system.