| The stability of the power grid is related to the development of all walks of life in China.In order to ensure the safety and reliability of the power grid and its equipment,it is necessary to regularly inspect the power lines to grasp the working conditions of the equipment and the environmental changes around the lines.The traditional power line inspection methods inevitably consume a lot of manpower,financial and material resources,and the inspection efficiency is low.With the development of remote sensing,the use of UAV inspection methods can avoid many drawbacks of traditional inspection methods,reduce the occurrence of missed inspections by manual inspection methods,reduce the time required to complete the line inspection task,reduce labor and financial expenditures for inspection power lines reduces the work.It is currently the most widely used method of power line inspection.In this paper,the vertical takeoff and landing fixed-wing UAV model cw20 is used to obtain the power line image by aerial photography.By extracting the power lines in the image,the purpose of using the UAV to patrol the power line is achieved.The main research contents include:1.Since the power line image taken by the UAV often has blur or noise,in order to improve the image processing efficiency and data accuracy,it is necessary to pre-process the image before analyzing the image,improve the image quality,attenuate background information and highlight target information.This paper first completes the preprocessing operations,including image graying,histogram equalization,filtering denoising,and so on.2.Edge detection is performed on the preprocessed image.Six edge detection algorithms are given in this paper: Roberts operator,Sobel operator,Prewitt operator,Laplacian operator,LOG operator,Canny operator.Edge detection of power line images is performed using these six operators.The experimental results show that: 1)Roberts operator has low edge positioning accuracy and can not suppress noise,which is suitable for low noise and steep images.2)The Sobel operator has low positioning accuracy and has a partial smoothing effect on noise,and is suitable for low-noise and gray-graded images.3)Prewitt operator edge positioning accuracy is not high,it has a certain smoothing effect on noise,and it’s suitable for gray gradient image.4)The Laplacian operator has accurate edge localization ability,is sensitive to noise and grayscale gradation,and can accurately locate stepped edge points.However,there are two-pixel boundaries,and the Laplacian operator is rarely used alone to detect image edges.5)The LOG operator has high positioning accuracy and is greatly affected by noise.It is suitable for images with no noise or low noise,and has good anti-interference.6)The Canny operator’s edge positioning has high precision,although it is affected by noise,but it has better background detection effect.The edge detection images of five kinds of operators are analyzed.The results show that the Canny operator has the best edge detection and enhanced edge integrity.3.After analyzing the development process of the straight line extraction,the power line extraction is performed on the image after the edge detection.In this paper,Radon transform and Hough transform are used to extract the power line.The image of the edge detection is extracted by Radon transform and Hough transform.The advantages and disadvantages of the two are given according to the experimental results.Since the Radon transform can accurately extract the power line,the start and end points of the power line cannot be determined,resulting in the power line running through the entire image;Hough transform can avoid similar problems,but it has a large amount of calculation and long running time;The power line is extracted by the Radon transform method based on Hough transform.The experimental results verify the accuracy of the method to extract the power line and avoid the disadvantages of using the Radon transform or Hough transform to extract the power line. |
PDF Full Text Request