Automatic Extraction And Matching Method Of Power Line In Complex Background Image Of Unmanned Aerial Vehicle

The power line is an important part of the transmission line,regular inspection of power lines is to ensure the normal operation of the transmission line,.Using the technology of unmanned aerial vehicle to replace the traditional method of power system maintenance is of great significance to the safe of power grid.Although there are a number of algorithms proposed by domestic and foreign scholars,the existing power line extraction methods are unable to handle the aerial image under complex background,and the accuracy of recognition ability of the power line is limited.On the other hand,after extracting the features of power line successfully,few research propose the method of 3D modeling using the result of extraction,especially bundled conductors.To solve these problems,this paper propose a new power line extraction method of aerial image in complex background.The paper use LSD(Line Segment Detector)line feature extraction algorithm,combined with the characteristics of power line gradient symmetry and the circularity of bundle conductors,to achieve the recognition and 3D modeling of the power line in aerial image,the research contents include:(1)To solve the problem of low recognition and accuracy rate in current method,this paper analysis and learn from the existing power line extraction algorithm of UAV(Unmanned Aerial Vehicle)aerial image proposed by domestic and foreign scholars.This paper proposed a power line recognition algorithm based on UAV aerial image,using LSD and gradient symmetry of power line.(2)After analysing the theory of bundled conductors in power transmission line and overhead line equation,taking advantage of geometric features of bundled conductors and line catenary equation,this paper propose a automatic matching algorithm for power line,in order to diagnosis or troubleshooting effectively on power line.Experiments prove that the proposed algorithm of power line has greatly improved the recognition accuracy for the UAV images with complex background,no matter images background is vegetation,bare land or road,the 3D model of power line finally get better.