Application Research Of Deep Learning In Infrared Diagnosis Of Voltage Heating Equipment

Infrared temperature measurement is widely used in power system,which is a non-contact and nondestructive measurement method.It can accurately measure the temperature and judge the operation status of power equipment.With the proposal and implementation of smart grid strategy,the combination of safe and stable operation of equipment and artificial intelligence is a key research direction.As the core technology of artificial intelligence,deep learning has been widely used in the field of electrical equipment fault diagnosis.It is combined with deep learning to find a breakthrough for the problem in infrared image analysis of voltage heating equipment,whose heating is insignificant and temperature width is narrow.The infrared image background of the equipment obtained from patrol inspection is complex.If a single feature such as artificial feature,insulator category and scene is used to identify and diagnose the equipment,it will not only get low recognition rate.but also be difficult to represent the target feature of the fault image.This paper presents an infrared diagnosis method of voltage heating equipment based on deep learning.Firstly,the infrared image is preprocessed,and the infrared image filtering based on adaptive mask can remove gaussian noise and pepper noise well.Then,the key area of the target equipment is obtained by establishing the target detection model of voltage heating equipment based on Centernet.On this basis,three typical features of the infrared image of voltage heating equipment are extracted,including color,wavelet and texture feature,and then the fault feature of three different attributes are fused through the serial fusion algorithm to fully extract the fault feature of infrared image of equipment,so as to replace the original image as the input of DBN,then use DBN to learn and train the samples,and finally classify the infrared images through the trained model to complete fault diagnosis.The experimental result shows that compared with support vector machine and traditional neural network method,DBN can achieve better result in infrared diagnosis of voltage heating equipment.Additionally,the fault diagnosis system is built through MATLAB.and the method proposed in this paper is embedded in it,which can achieve infrared diagnosis well.At the same time,the infrared diagnosis method based on deep learning proposed in this paper is combined with auxiliary ways such as live detection,power failuredetection and disassembly test to further diagnose the fault of voltage heating equipment.