Research And Implementation Of Substation Video Analysis System

In recent years,with the rapid development of power information technology,artificial intelligence technology and power grid scale,the operation and management mode of power enterprises has undergone significant changes.The current duty mode is to transmit all kinds of real-time data of substations to dispatching centers,which reduces the influence of human factors.However,it still needs to be judged by human analysis of surveillance video whether illegal intrusion occurs in substations and whether work violations occur.This duty mode,which relies on staff to view video day and night,consumes a lot of manpower.It can be seen that transforming the existing substation duty mode and realizing more intelligent management mode have become the inevitable trend of substation operation and management development.With the rise of in-depth learning,object detection algorithm based on in-depth learning can recognize objects efficiently and accurately.This ability is very suitable for the requirements of current substation duty mode.Therefore,in view of the outdoor working environment of substations,a convolutional neural network applied to video analysis of substations is designed and implemented,and safety clothing,safety helmet,insulating gloves and work are completed.Identification of seven kinds of objects such as certificates.With the help of identification results,the problems of illegal intrusion,no safety helmet and other items,no operation in the prescribed time and place are analyzed and alerted.The system is centered on target recognition and video analysis.It is mainly divided into five modules:user registration and login,operation safety rules,object recognition,violation analysis,log management and system management.Operational safety rules are the specific requirements of administrators for a certain operation,including the setting of time,place,number of people,safety helmet and other items.Violation analysis module reads video images of substations,identifies some targets,identifies specific employees,and compares them with operation safety rules to get the analysis results.