The Design Of Remote Video Monitoring System Based On Event Driven

Coalbed Methane is a kind of clean energy,which is abundant in our country,so China is putting more and more attention on the exploitation of coalbed methane.However,coalbed methane is a kind of flammable and explosive gas at same time,so explosion accidents may happen in the process of mining.Most of the methane wells lie in remote mountains and very scattered,so manual inspection is quite difficult and time-consuming and with some security risks.At the same time the cable video monitoring costs too much to lay the cable.To solve these problems,this paper designs a remote video monitoring system based on event driven.Firstly,this paper summarizes the traditional moving target detection algorithm and analyzes the frame difference,single Gauss,mixed Gauss background model and codebook algorithm.For pump reciprocating motion exists in the scene monitoring,puts forward a method of building a background model based on region segmentation and motioning moving target based on different region.Firstly,the background is divided into static background region and dynamic background region by the background region segmentation algorithm.And then for the static background region,the paper uses three frame difference algorithm for moving target detection and uses the improved codebook method to detect the moving target for the dynamic background region.The simulation results show that the algorithm proposed in this paper can distinguish the complete trajectory of pump through the online learning,and then the dynamic background region and static background region will be separated.So we can choose a simple static detection or dynamic combined with static region detection method based on the actual industrial application.The simulation results also show that compared with the traditional algorithm,the accuracy of the algorithm is greatly improved,and the time of target recognition is reduced effectively.Next,this paper introduces the hardware structure and program design of the capture client,and designs the networking method of the system.The video capture client's hardware is mainly composed of solar power,wireless router,intelligent camera and the core component-Exynos4412.The client program includes a number of threads,such as the watchdog thread,the monitoring thread,the acquisition thread,the sending thread and so on.The system network is built by VPN based on the L2 TP protocol.The server is programmed by the Windows platform using MFC,which is consisted of five modules,mainly for image decoding,display,storage,logging and other functions,including the function that send commands to client(such as ask for a picture,remote reboot system),so it has a good human-computer interaction.At last,this paper gives an overall test of the whole monitoring system,including the strength of network on field,network delay,the accuracy and real-time capability of the system,the stability and reliability of the client hardware(including the waterproof and dustproof and temperature test)and so on.After the test,the system is installed and applicated in the field.In the actual application process,we also found some problems such as the false alarm detection in night.etc.And these problems provide the direction for further improvement of the system.