Research On High Voltage Power Lines Autonomous Inspection And Fault Recognition Algorithm Based On UAV

High voltage power lines are an important part of the power system.Regular inspections of high voltage power lines can ensure the safe and stable operation of the power grid.At present,with the development of artificial intelligence technology,it is an inevitable trend to use UAV to conduct high voltage power lines inspections.To realize the autonomous inspection of high voltage power lines based on UAV,it is necessary to automatically recognise and extract the high voltage power lines,and then calculate the forward direction according to the recognised high voltage power lines,and perform the fault recognition of the insulator during the inspection to improve the efficiency of the inspection and reduce the case that staffs go to the site to conduct inspection.Aiming at the difficulties,this paper proposes a high voltage power lines recognition method based on improved LSD line segment detection algonithm combined with high voltage power lines object features,and some fault recognition methods about the fault of string breakage,snow coverjing and iceing of the insulator based on insulator multi-feature information is proposed.Finally,the related algorithms are verified by MATLAB and Microsoft Visual Studio 2010 simulation.The main work is as follows:(1)According to the characteristics of the automatic inspection of the high voltage lines based on UAV,the design of the autonomous inspection system for the high voltage lines based on UAV is studied.Design was performed for the four subsystems including the body subsystem.The working principle and working flow of autonomous inspection of high voltage lines based on UAV are suggested.The common electrical equipment faults are analyzed.Two problems that need to be solved in this paper are studied,the recongnition of high voltage lines and the automatic recognition of insulator faults.(2)According to the characteristics of the UAV inspection images,the image preprocessing algorithm is studied.Due to the change of external illumination conditions,it is easy to result in overexposure or underexposure of the UAV inspection images.To solve this problem,this paper proposes an optical rectification algorithm based on gamma transformation,which can effectively improve the image quality degradation of the UAV inspection images.Simultaneously,aiming at the noise interference in the UAV inspection images,a recognition algorithm based on wavelet transform is used,which greatly improves the quality of image restoration.On this basis,we will further study the autonomous inspection of UAV high voltage lines,most of which are carried out in a complex environment,which leads to the complex background of the UAV inspection images.For the problem that the power line is difficult to recognise and many interfering objects,the power line is improved and analyzed by improving the traditional LSD algorithm.Through the improvement of the traditional LSD algorithm and the analysis of the high voltage power lines features,a automatic recognition and extraction algorithm combined with the improved LSD algorithm and high voltage power lines features is proposed,which can accurately and qulickly recognise and extract high voltage power lines,which provides a solid foundation for UAV autonomous inspection on high voltage lines.(3)In order to realize fault recognition of the insulator during the autonomous inspection based on UVA,the essential features of the insulator are analyzed,and the description methods of insulator are obtained by using texture feature,color feature,shape feature,invariant moment feature and wavelet coefficient feature.Aiming at the fault of string breakage,snow coverjing and icing of the insulator,the corresponding fault recognition algorithms are proposed and verified by simulation.