Research On Insulator Identification And Spectral Defect Detection Technology Of Transmission Lines

Insulators play a key role in electrical insulation and line support in the transmission line.Once the insulators are defective,the entire transmission line will be faulty,and even a large area of power outage will occur.Therefore,timely detection of the insulator operation status is especially important for the stable operation of the transmission line.Traditional manual observation and detection methods are subject to geographical restrictions,and ultrasonic detection methods have short application distances.With the maturity of image processing technology,the introduction of "smart grid" and the rapid development of the UAV industry,the UAV aerial photography detection method makes up for the shortcomings of the traditional methods and gradually becomes the mainstream detection method.Using the image to detect the operating state of the insulator,the insulator is first identified from the image.In this paper,the identification method of the insulator is studied in detail,and the detection and classification methods of two common defects for different spectra are designed based on the identification.The paper firstly uses the SVM classifier to identify the insulator based on the Hu invariant moment and HOG features of the image.The experimental results show that this method has a high accuracy in the case of single background,but the accuracy will be greatly reduced in the case of complex background.The aerial image of UAV usually has complex background,so the applicability of this method is not strong.The existing method of red-blue difference graying and clustering can segment and recognize insulators with complex image background,but there is too much background in the segmentation result and the effect is not ideal when the composite insulator image is grayed out.Therefore,this paper proposes to use the Otsu algorithm to preprocess the image first,and then we respectively perform the red-blue difference graying and red-green difference graying on the preprocessing result.Then the K-means clustering algorithm is used to segment the gray images.Finally,based on the spatial configuration consistency of insulator,the projection feature is used to identify the insulator.The experimental results show that the improved method can segment and identify the insulator image of the transmission line more effectively.On the basis of identifying insulators,a method based on improved UL-PCNN and regional pixel ratio is proposed to detect and classify two different spectral defects of"self-shattering" and "corona discharge".First,the red-blue difference gray image is used to replace the traditional weighted gray image as the external excitation of the UL-PCNN fine segmentation model,and then the detection window is drawn for the segmentation result,and the defect region is detected by comparing the number of white pixels of each window.Finally calculate the proportion of white spots in the defect area to achieve defect classification.