Research On The Application Of Image Processing Technology In UAV Power Line Inspection

China is a large power consumption country.The safe and stable operation of power grid is closely related to the development of national economy.As the main part of power grid,power line is the basis of stable and safe operation of power system and power grid.The power lines used in long-distance power transmission and short-distance power distribution in China are basically steel-core aluminum strands,and steel-core aluminum strands will produce different forms of faults under the influence of extreme weather,among which foreign body attachment and power line stock breaking are the most common.Therefore,regular inspection of power lines is the key to ensure stable and safe power supply and ensure the national economy.The existing manual inspection method in China is difficult to meet the needs of large and complex transmission network maintenance in China.It is urgent to explore and apply new inspection methods.UAV inspection has become an urgent need for the development of power system.With the development of power system automation and smart grid,image processing technology and computer vision technology are increasingly used in smart grid inspection.Based on the power line images captured during UAV inspection,this paper uses image processing method to realize power line identification and fault detection.The specific work is as follows:Firstly,in view of the characteristics of complex background and high random noise of power line image,this paper preprocesses the power line image,and uses image graying,histogram equalization and median filtering to improve the quality of power line image.An improved Canny operator is used to detect the edge of the image.The traditional edge detection operator is improved in the two directions of gradient amplitude calculation and double threshold determination,which improves the anti-noise ability of the traditional Canny operator and enhances the adaptive ability of the Canny operator.Secondly,based on the preprocessed image,the power line is identified and extracted according to the linear characteristics of the power line.In this paper,three commonly used line detection algorithms,Radon transform,LSD algorithm and Hough transform,are analyzed.An improved Hough algorithm is used for power line recognition.This method combines the results of probabilistic Hough transform and standard Hough transform,which effectively avoids the repeated recognition of power line and can accurately label the endpoint of power line in the image.Then through the regional segmentation algorithm,the power line as a whole as a connected region is segmented from the image,which lays the foundation for fault detection.Finally,this paper gives effective detection algorithms for two common faults.For power line foreign body,this paper adopts a foreign body detection algorithm based on contour detection.Firstly,the image is binarized,and the projection method is used to estimate the suspected foreign body area.Then,based on multi-group morphological operation,a power line foreign body detection algorithm based on contour recognition and screening is given to realize the detection of power line foreign body,and the foreign body is cut and extracted.For the power line break,this paper detects the fork of the transformation component of the break detection.By skeletonizing the power line,the identification criteria of power line endpoint,line point,intersection point and isolated point are defined,and the detection of power line break is realized.In addition,this paper also designs an algorithm based on Harris corner detection and corner screening,which effectively eliminates the interference of power line virtual cross on power line break detection.