Design And Implementation Of Power Grid Video Monitoring Behavior Analysis System

With the rapid development of national power informatization,the demand for information collection and analysis is increasing day by day.Because there are realistic and potential various risk factors in power production and operations,traditional video surveillance systems The function is often relatively simple.It is that the surveillance camera transmits the data collected by the cameras distributed in each surveillance scene to the surveillance room,and then realizes the surveillance through manual participation.However,in the process of power production,the surveillance personnel needs to be uninterrupted.Go to work.For the monitoring staff,the labor intensity is very large,and it is also easy to cause excessive fatigue of the staff,and it is difficult to ensure the accuracy and efficiency of the work.Therefore,if the computer is used to replace the person to carry out the entire process,the automatic monitoring becomes It is very necessary.First of all,the method of detecting moving objects and the background model are selected.In this paper,the inter-frame difference method and the mixed Gaussian background model are used to detect moving objects.Since the power monitoring video is not only indoors,but also outdoor,there are Light also has the influence of various natural environments,and since real-time monitoring of moving targets is required,a method with good real-time performance is needed for detection.The method chosen in this paper is not particularly sensitive to the influence of noise such as light,has good anti-interference ability,and the background model can be obtained better when there is a moving background in the background.Then in this paper,the method of combining the inter-frame difference method and the mixed Gaussian background model is used to detect the target.First,we use the mixed Gaussian background model to preprocess the target video,separate the foreground target and the background model,and then use The inter-frame difference method video is processed to obtain a more accurate model of the contour of the moving target.This paper proposes an improved mixed Gaussian model.Because the algorithm of the mixed Gaussian model is more complex,the calculation speed is slower and cannot be achieved very well.Real-time effects.Therefore,in the improvement,the calculation amount of the mixed Gaussian background model is mainly reduced,and the parts that have been determined as the foreground and the background do not need to be repeatedly determined to improve the calculation speed of the algorithm and meet the requirements of the system.In this paper,the main research object is grid monitoring video,so the final result is more targeted.Finally,the grid video surveillance behavior analysis system is designed.The grid video surveillance behavior analysis system is mainly composed of three aspects: video input module,video display module,video analysis module,video training and test module.Each module has different functions.The video analysis module is the core of the system.It is mainly responsible for reading the video from the video recording module and analyzing the video.Finally,it conducts actions that violate the regulations or have hidden dangers.Alarm reminder.The main work in this article is the video analysis module,which mainly deals with video analysis,namely moving target detection,moving target tracking,and moving target behavior analysis.And made a detailed introduction to the detection of moving targets.