Design And Implementation Of Substations Based On Intelligent Video Analysis

Substation charged with the core task in the power system, it has an important impact in its supply and distribution area. Based on intelligent video analysis, management of substation forms are extensive researched. By this intelligent management model, network managers can easily achieve remote monitoring, and can more quickly find variety security risks, then timely process the problems, which eliminated a lot of work of maintenance and troubleshooting, effectively reduced staff’s working time and working intensity. This design of substation intelligent monitoring system has a very important significance to build a visualization unified intelligent management of power system.In response to these needs, we designed the substation monitoring system base d on intelligent video analysis. This system intelligent video analysis technology into the unified power grid video surveillance platform, which achieved intelligent troubleshooting and automatic alarm function, and can work 24 hours a day, at the same time, it can filter and differentiation monitoring data, which eased the storage pressure and manager ’s workload, improved the overall efficiency and reliability of the monitoring system.Substation intelligent video analysis is mainly focused on equipment condition monitoring, intrusion and abnormal behavior detection, violation detection, video quality detection modules. We conducted a comprehensive analysis from the device state to the surrounding environment. And designed the hardware and software architecture of the system, then introduced each constituent elements in the functional modules, and functions can be achieved in each module. Meanwhile, directed to the most important security aspects of substation, the equipment condition monitoring and intrusion detection, the paper focuses on the substation digital instrument recognition and motion targets tracking.In order to achieve digital instrument recognition, a multi-feature multi-classifier fusion recognition algorithm proposed in this paper, the method combined several groups of typical characteristics which have high degree of distinction and simple calculation, and then via the minimum distance classifier and the nearest neighbor classifier achieve twice classification. By combining multiple features, we formed a recognizable characteristic factor, and via coarse and fine classifier two layers of screening, it is able to guarantee the accuracy of the identification system and simultaneously improve recognition speed. Test showed that: the recognition rate up to 99.03%, and the system identification time is less than 300 ms, which achieved the speed and accuracy requirements.In order to solve the problem of substation intrusion detection, this paper studied the moving target tracking algorithm. The accuracy of object tracking algorithm determines the alarm correct or not, this article researched the Mean Shift algorithm and Cam Shift algorithm theory and the process of achieve moving target tracking, then based on this algorithm we use VC ++ and OpenCV achieved moving target tracking program, experiments show that the algorithm tracking accuracy satisfy the actual demand.