Research On Electrical Equipment Infrared Image Fault Area Segmentation And Fault Diagnosis

With the continuous advancement of my country’s smart grid construction,my country’s electric power industry has made unprecedented progress.In order to ensure the safety and reliability of the normal operation of the power grid,maintenance personnel usually use testing instruments related to fault diagnosis to conduct regular inspections and fault detection of electrical equipment in operation.However,maintenance personnel are required to analyze and judge the surface temperature distribution of the equipment,and the detection efficiency is relatively low.With the development of computer technology and image processing technology,automatic substation inspection robots have been widely used in the power industry and have become an important means of monitoring the status of electrical equipment.This paper takes the collected infrared temperature measurement images of electrical equipment as the research object.Through image processing,super pixel segmentation,feature extraction,classification recognition and relative temperature difference method,the purpose of automatic detection and diagnosis of electrical equipment is realized.In order to reduce the labor intensity of the inspectors and the dependence on the experience level,discover problems in time to avoid losses,and ensure the safe and stable operation of the power grid.The main topics of this paper are as follows:(1)As the resolution of infrared images collected by modern infrared detection equipment is gradually increasing,traditional pixel level image segmentation technology is difficult to meet real-time requirements.Therefore,a super pixel segmentation technology that has emerged in recent years is studied.Compared with common image segmentation technologies,super pixel segmentation technology can not only retain the local information of the image,but also increase the operation speed of subsequent algorithms.The infrared image segmentation experiment of electrical equipment is carried out on the super pixel algorithm through the super pixel evaluation standard,and the performance of different super pixel algorithms is analyzed,so as to lay the foundation for the follow up research of this paper.(2)In order to realize the automatic processing of the infrared image of electrical equipment and improve the segmentation accuracy of the equipment fault area in the infrared image,In the process of generating super pixels by SLIC pre-segmentation,there will be problems such as fragmentation of super pixel regions and insufficient accuracy of super pixel boundary segmentation.An improved SLIC algorithm for electrical equipment fault region segmentation method is proposed.The infrared image is preprocessed by using a guided filter,and then the brightness similarity limit condition is increased in the SLIC super pixel iteration process,And the generated super pixel is matched with the color value of the seed point,and finally the segmentation and marking of the fault area of the electrical equipment is realized by automatically setting the hue threshold.(3)Propose a recognition and classification algorithm of electrical equipment infrared image equipment based on Histogram of Oriented Gradient feature extraction and support vector machine.The gradient of the image and the distribution of the edge direction density can better describe the shape of the local area,and realize the identification and classification of electrical equipment.Combined with the characteristics of the thermal state of the equipment and the corresponding criteria,the automatic detection of thermal faults in electrical equipment is realized.Design a thermal fault detection and diagnosis system for electrical equipment,and conduct experiments on common electrical equipment such as disconnector,wiring connectors and high-voltage bushings.Through the analysis of the experimental results,the fault identification and diagnosis method of electrical equipment based on infrared images proposed in this paper can determine the degree of thermal fault defects.The validity and reliability of the algorithm in this paper are verified.