Research On Application Of Video Recognition Technology In Substation

Video monitoring system is an auxiliary monitoring means for operation and maintenance of power network.With the development of video surveillance technology and intelligent analysis technology,video surveillance business is developing towards intelligent,high definition and networking.With the continuous develop ment of power industry,the power sector built high voltage power supply and transmission network,leading to the traditional monitoring technology has been unable to carry out dynamic monitoring of such a large-scale power grid.The subjective judgment of the operation personnel and the way to monitor the large number of digital instruments,not only can't guarantee the accuracy,but also a great waste of manpower and material resources.At the same time,the security problem of substation is also need to focus on,any minute security issues is likely to cause the accident,if there are people into the important regional or intentional destruction,this would not only cause huge economic losses,but also lead to the normal power supply,the impact of people's lives and other issues.Therefore,this paper focuses on the problem of intelligent video surveillance system,mainly carries out on the digital instruments and important regional personnel to the two aspects of the invasion of the study.Firstly,the current research status of video surveillance,digital instruments and moving target detection is analyzed,and the main technology of current video recognition is analyzed.Secondly,through the study of the current mature digital instrument recognition method,this paper proposes a digital recognition method of intersection based on feature extraction,the method is to improve and optimize the sewing method,before character recognition,we set up some restrictions on the number of the segmentation,mainly from the area of the rectangular area,the height difference and the vertical coordinates of the three aspects of the limit,to prevent judging the noise area to be a number incorrectly.The recognition feature is constructed by the intersection number of lines and characters,and recognition of digital 0-9 is carried out.The algorithm can greatly reduce the computational complexity and has high application value in the condition of ensuring the accuracy of the identification and anti-jamming.Finally,this paper studies the moving target detection methods of the current mainstream,including optical flow method,background subtraction method and frame difference method.By comparing the advantages and disadvantages of the above methods,this paper proposes a method of background subtraction and the two frame difference of moving target detection method.This method is to do and operation to difference maps of background subtraction method and image sequence difference method,and the image sequence difference method is improved,assimilated into the influence of illumination changes.The method can avoid "double image" phenomenon caused by image sequence difference method,and avoid the light mutation to be classified as moving target that caused by the background subtraction method.It improves the recognition accuracy,and it is relatively simple to achieve,can meet the real-time requirements of intelligent monitoring system.