Study On Malfunction Image Analysis In Inspection Of Substations

Traditional inspection methods are difficult to be comprehensive and accurate,and are prone to missed inspections,false inspections,etc.,and through the analysis of the existing substation inspection technology and development needs,the use of inspection robots as carriers to various electrical substations The inspection of equipment provides a reliable basis for the safe operation of substations and is also the development trend of smart grids.This article is based on image problems collected during inspection of substations,and mainly studies the methods of image analysis and identification.Firstly,the basic principle of infrared image temperature measurement is explained.Then,the detection process of electrical equipment when a thermal fault occurs is explained.Then,the relationship between device heat generation and equipment failure is analyzed to find out the abnormal temperature distribution on the infrared image and realize electrical heating.The temperature difference of the infrared image of the device is detected.Due to the disadvantages such as poor contrast and low signal-to-noise ratio in infrared images,the morphological color gradient component of the image is extracted,the Hessian matrix of the image is used to construct a multi-scale filter,the image is filtered and enhanced,and the filtered image is enhanced.The watershed segmentation is performed and the segmented region growth algorithm is used to merge the over-segmented regions to segment the faults from the background image.In order to more accurately identify the type of faulty device,an improved SIFT algorithm can be used to register the light image.The algorithm mainly includes the use of improved SIFT algorithm for feature detection,bidirectional matching algorithm for feature rough matching,and RANSAC algorithm.The matching point pair is purified and the fitness matrix is calculated.Finally,the image is registered and the faulty device type is accurately identified.