Research On Infrared Image Processing Of Electrical Equipment In Substations

With the rapid development of the power industry,the capacity and voltage level of the power grid have also been continuously improved.In order to ensure the reliability and safety of the power supply,it is particularly important to monitor the temperature of the electrical equipment in operation and automatically diagnose the fault.At present,infrared technology has been widely used in power equipment testing in foreign countries.However,the use of handheld infrared cameras is still widely used in China to conduct inspections on power equipment.The efficiency is low and it is prone to missed inspections.In view of the shortcomings of traditional detection methods,this paper designs a fault detection system for substation equipment based on infrared thermal imaging technology,which monitors the running status of electrical equipment to ensure reliable power supply.main tasks as follows:(1)Firstly,the noise of interference infrared imaging is analyzed.Comparing the mean filtering,median filtering and transform domain filtering effects,it is found that the traditional filtering method will lead to denoising image blur and edge information loss.Under this premise,it is proposed to use non-subsampled Contourlet changes for filtering.By comparing the advantages and disadvantages of soft threshold,hard threshold and compromise threshold,it is proposed to use the improved adaptive threshold to judge the NSCT coefficient value.With the help of MATLAB simulation analysis,the feasibility of denoising by non-subsampled Contourlet transform is theoretically verified.(2)Based on the analysis of image segmentation principle and evaluation criteria,an image segmentation method based on pulse coupled neural network is proposed.Compared with the traditional pulse-coupled neural network,the parameters are complex and difficult to determine.A simplified pulse-coupled neural network is used as the segmentation.The algorithm is combined with the cross entropy to set the number of iterations,the link coefficient matrix and the modulation coupling coefficient are improved,and the experimental simulation is carried out.The experimental results show that the improved algorithm can segment the hot part of the image and improve the detection accuracy.(3)According to the characteristics of infrared image of substation power equipment,feature extraction is performed by Hu invariant moment.The invariant moments extracted from the rotation,mirroring and shrinking deformation of the same image show that the lower moments have good invariance.On the basis of identification,the fault type of power equipment is analyzed,the fault diagnosis rules are designed,and the temperature information in the infrared image is extracted.The infrared image fault diagnosis system based on surface temperature judgment method is realized.