Research On Algorithm Of Early Warning System Based On Intelligent Video For Transmission Line Breaking

With the rapid development of society,Chinese electricity consumption is rising,and more and more power grid accidents caused by external damage of transmission lines will happen.The traditional early warning by manpower not only wastes manpower and material resources,but also has high rate of missing and late warning.The existing automatic early warning device is easy to be interfered and has high rate of false warning.As a result,it is of great significance to study an automatic,stable and real-time early warning system against external damage.In recent years,with the rise of computer vision technology in China,researchers began to study the early warning system based on intelligent video against external damage of transmission lines.In the external damage factors,the incidence rate of the collision between boom and lines is high and the harm is great.Therefore,this paper studies the early warning system of the boom intrusion into the transmission lines.In this paper,through the research of image preprocessing,feature extraction,moving object detection,recognition,tracking and binocular stereo vision,and the requirements analysis of early warning system,then,the functional design is proposed.By studying the design and implementation of the existing algorithm,and based on the experimental analysis of the shortcomings of the algorithm,the following improved algorithm is proposed,(1)The detection algorithm of transmission line based on Hough Line is proposed,which lays a good foundation for the early warning algorithm of distance measurement between transmission line and boom.(2)Through the study of optical flow method,frame difference method and background difference algorithm,the advantages and disadvantages of the algorithm are compared.Basing on analysis,a target detection algorithm combining improved five frame difference and single Gaussian background modeling method is proposed.(3)Through in-depth study of meanshift tracking algorithm and boom geometric features,a universal rectangular box based on SIFT feature points is proposed as tracking frame,and CAMSHIFT tracking algorithm is improved.Thus,the problem that the existing tracking frame contains a large number of background pixels and the scale direction of rectangular box cannot be adjusted adaptively with the boom.(4)Based on the in-depth study of binocular stereo theory and the characteristics of the collision between boom and line,a distance measurement algorithm of transmission line and boom based on stereo correction and fusion of common vertical distance of different plane lines and point to line distance is proposed,which solves the problems of large ranging error and multiple contact blind areas of existing system ranging early warning algorithms.By conducting experiments,it is proved that the transmission line detection algorithm can detect transmission lines out of doors and be used for ranging and early warning.The improved moving object detection algorithm can not only improve the detection efficiency but also the detection result.The tracking box of the improved tracking algorithm can adjust adaptively with the scale and direction of the boom,so the tracking accuracy will be improved.The improved ranging algorithm not only improves the ranging accuracy and warning accuracy,but also greatly reduces the rate of missing warning.By comparing with the existing system,the system performance verify the practicability and superiority of the system.