Design And Key Technology Of Integrated Video Surveillance System For Railway Stations And Depots

The integrated video surveillance system for railway is a key part to ensure the safe operation of railway.In the daily transportation of railway stations and depots,there are some characteristics such as scattered workplaces,more repetitive labor,different labor intensity and different shift system,etc.,which have great challenges to the daily transportation safety and management.If intelligent video image analysis technology is applied to railway department daily production activity,can effectively improve the efficiency of railway transport production,ensure safety and reduce the cost of management,and can realize equipment to replace the unpopular remote work post,and take the management from traditional manual released in daily management.The design of railway station and depot integrated video monitoring system with comprehensive safety monitoring,dispatching and comprehensive analysis functions can systematically and scientifically manage the operation site,enhance the supervision of railway departments on various operation processes and the speed of emergency treatment of sudden safety accidents,and improve the safety of railway transportation.This thesis studies and masters the development and application of video surveillance technology at home and abroad and its application in the railway industry.The thesis collects many daily actual demand data of railway stations and depots.Then it establishes the top-level architecture of railway stations and depots` integrated video surveillance system.As well as,the thesis finds out the key technical solutions for the system to realize the functions of security monitoring,production scheduling and comprehensive analysis.In terms of safety monitoring,automatic alarm,special scene staring and monitoring linkage functions are designed and realized.In view of the special scene staring control,combining with the station cargo security inspection,monitoring linkage technology,this thesis applicates the discriminant method based on the frame difference method for the flow direction of the crowd in and out of the station,the convolutional neural network based on the face recognition and inspection.The technical scheme of station cargo safety inspection is designed,and the safety monitoring linkage mechanism is established.In terms of production and service management and control,model optimization research is carried out from three aspects of passenger transport,freight transport and railway traffic operation.Focusing on railway intelligent cargo inspection,an intelligent recognition model of cargo inspection image based on improved SSD/YOLO deep learning is proposed to realize the optimization of cargo inspection image recognition.In terms of video comprehensive analysis,this thesis divides system monitoring protection zones,adopts "the front + back end" video analysis method.In the front analysis,a comprehensive analysis of the front-end integrated with safety intrusion detection method based on Gaussian mixture model,based on pixel statistics of crowd density estimation method,are used.And in the information acquisition,transmission,and storage stage,network and information security protection is guaranteed.Finally,taking GZN railway station as an example.The top-level architecture of the railway stations and depots` integrated video surveillance and the functional technologies of safety monitoring,production scheduling and comprehensive analysis are applied.The application effects are evaluated,and the system features are summarized.