Research On The State Diagnosis Technology Of Live Operation Of Multiple Circuits In Same Tower Based On Image Recognition

With the continuous expansion of power grid capacity and coverage,the continuous improvement of transmission voltage level,the complexity of power grid architecture is getting higher and higher,and the requirements for power supply quality are getting higher and higher,and grid security plays a pivotal role.Highvoltage live maintenance technology refers to a technical means to carry out live detection,analysis,diagnosis and maintenance of high-voltage lines and equipment within the specified maintenance cycle,so as to ensure the reliability of power supply.It has the advantages of safety and efficiency,green environmental protection,does not affect the operation of the power grid,etc.,can not only realize equipment maintenance,elimination,but also maintain continuous and uninterrupted power supply,economic and social benefits are remarkable.It is widely used in the maintenance,transformation and upgrading of the power system,which provides a guarantee for the reliable and stable power supply of the power grid.The multi-loop line of the same tower is to set up multiple transmission lines on an iron tower,for transmission in different directions,but due to the centralized layout of transmission lines,it may lead to discharge accidents or electric shock accidents and threaten the life safety of live workers,so it is necessary to monitor the distance between live workers and substation equipment in real time and accurately when working with electricity.In this thesis,the live operation status diagnosis technology of the same tower multi-loop line based on image recognition is studied,and the main work is as follows:(1)In this thesis,the picture collected by the camera is first extracted by the mixed Gaussian background difference method,and the algorithm solves the problem of missing edge of the moving target.Morphology was used to fill the extracted human voids and eliminate spot noise.(2)In this thesis,an algorithm combining particle filter and Kalman filter is proposed to track moving targets.The tracking algorithm consists of two stages:firstly,the color histogram is used to extract the features to obtain the search area,and secondly,the area is filtered to find the location with the highest degree of matching,which improves the power grid security management level.(3)The principle of monocular vision system is studied in depth,the camera is calibrated according to the principle of monocular vision calibration,the parameters of the camera are obtained,and the actual distance represented by the unit pixel of the picture is calculated by the conversion between coordinate systems.Calculate the actual distance between the operator and the insulation clamp,and then compare it with the safety distance evaluation index to determine whether the live working distance meets the standard,so as to improve the safety of the live operation.