Implementation Of Intelligent Video Server Interface Based On ONVIF Protocol And Human-Machine Interface Development

The main topic of this thesis is based on the intelligent video server interface of ONVIF protocol.On the ONVIF network video monitoring platform based on C/S architecture,the function service of ONVIF protocol is researched,including device discovery,media streaming service,event subscription service and embedded intelligent video Algorithm video analysis service,we design the standard common interface of server-side intelligent video analysis,to analyze the original video image,and send web services to the PC client to achieve intelligent monitoring,which has been a bridge between the camera and outside world.The versatility of the ONVIF protocol also makes the ONVIF video server becoming an independent,cross-platform intelligent analysis network server.The main work of this thesis is summarized as follows:(1)The thesis uses gSOAP toolkit to compile the development of ONVIF Web services document WSDL,to generate ONVIF protocol's C/S framework for the research and complete the SOAP communication between the server and the client for the implementation of the the relevant function services.(2)Then,the thesis designs and implement ONVIF protocol function services,including device discovery,media service,PTZ control,event subscription and video analysis.The device discovery is based on the WS-Discovery standard,so that the server can be actively detected by the client.Media streaming service is divided into video codec,format conversion module and real-time video streaming module.Event subscriptions are based primarily on WS-Notification specifications to enable event to subscribe and alarm.Video analysis services include video rule definition and analysis,embedded in various types of video algorithms to achieve a variety of video analysis services.(3)Implementation of video analysis interface.The internal video analysis algorithm is invoked using the ONVIF service interface layer and the relevant video analysis engine is configured.In this thesis,a moving object detection algorithm is used to study the motion target detection algorithm.The hybrid Gaussian algorithm is used to establish the background model for the image sequence.The background reduction method and the opencv basic bottom function are used to realize the motion target detection function.(4)Thesis constructs the plat of network intelligent video surveillance HMI.The user can define the monitoring area and subscribe to the relevant video analysis function.The video analysis server interface carries out the analysis and processing of the original video in the monitoring area,and realizes the intelligent monitoring when the monitoring value reaches the threshold.In addition,the client can obtain video analysis results,including the type of events and time,test results,and capture the target location parameters.