State Detection Of Electrical Equipment Based On Infrared Thermal Imaging Technology

With the rapid development of national modernization,the safe and stable operation of substation electrical equipment plays a very important role in social stability and economic development.The traditional fault detection methods mainly rely on manual judgment,which requires the technicians to have very rich experience and low detection efficiency.With the rapid development of artificial intelligence technology,the use of intelligent inspection robot equipped with infrared equipment for substation inspection is becoming one of the mainstream inspection methods.In order to solve the problem of accurate thermal fault detection of high-voltage electrical equipment in substation,and further improve the safety,stability and intelligent level of substation,this paper explores and realizes the state detection of electrical equipment in infrared image based on convolution neural network,which lays a foundation for automatic fault diagnosis of electrical equipment.Firstly,this paper constructs the Faster RCNN network model.Considering the poor quality of infrared image and the fact that it contains many kinds of devices,this paper selects vgg-16 as the basic feature extraction network to extract the features of all kinds of electrical devices in infrared image.In addition,this paper uses Faster RCNN target detection algorithm to classify and accurately locate a variety of electrical equipment and parts in the infrared image.The overall recognition accuracy can reach 83.1%,and good application results have been achieved.In the process of practical application,some parts of different equipment are easy to lead to recognition errors and high error rate due to similar appearance,which brings great trouble for subsequent fault diagnosis.In order to further shorten the time required and improve the accuracy of equipment location identification,this paper improves Faster RCNN by using NMS ideas.By calculating the area overlap ratio,the correlation between equipment area recommendations and equipment location area recommendations is established to carry out category correction.The recognition accuracy of the improved fast RCNN neural network model is 92.8%,which is 9.7% higher than that before.Then,in order to realize the real-time state detection and automatic fault diagnosis of electrical equipment,this paper designs and develops a set of infrared image intelligent analysis and diagnosis system based on Web,and applies the improved fast RCNN infrared detection model to the actual scene.Through the intelligent analysis and artificial assistant processing of the received infrared image,the state detection and fault diagnosis of the electrical equipment in the infrared image can be realized.