Fault Analysis Of Substation Equipment Based On Multi-objective Evolutionary Clustering For Infrared Image Segmentation

In recent years,image processing technology has developed rapidly.In order to ensure the safe and efficient operation of electrical equipment in substations,image processing technology and infrared temperature measurement technology have been introduced into the field of automatic detection of electrical equipment faults,which makes the daily operation and maintenance of the substation developed in the direction of automatic diagnosis.The automatic detection technology of electrical equipment faults can not only greatly reduce the experience and professional knowledge of daily inspection personnel,but also reduce the losses caused by thermal faults to substations.Image segmentation is an important link in the whole system,which provides strong support for understanding and recognizing the target state.In this context,this paper studies the denoising and segmentation of infrared images of substation electrical equipment,and applies it to the automatic fault detection of electrical equipment.This paper introduces the basic imaging principle of infrared image and analyzes the types of noise that may exist in the infrared image and the corresponding causes.On this basis,by analyzing several traditional image denoising methods,an adaptive mask image denoising algorithm is proposed.Compared with the traditional masking algorithm,this paper considers the similarity between each neighborhood pixel and the center pixel.In order to improve the filtering effect,the effect of each neighborhood pixel on the central pixel is adjusted by introducing the adaptive weight based on the similarity.Through the Matlab platform simulation and compared with other traditional algorithms,the experimental results show that the proposed algorithm has good denoising ability,robustness and wide application range,both for Gaussian noise and salt and pepper noise.Aimed at the segmentation of infrared images in substation electrical equipment,this paper proposes a two-layer framework that divides image segmentation into two steps.In the first step,considering the particularity of the inhomogeneity pixels,the original image is sampled by the sampling method based on the inhomogeneity measure,and the multi-object problem is decomposed into multiple single-object sub-problems by Chebyshev decomposition strategy.Then,using the evolutionary algorithm to obtain a trade-off solution based on the optimal solution of each sub-problem and its corresponding weight vector.The second step is to use the trade-off solution and the weight vector in the previous step to perform image segmentation.After setting the optimal parameters through the parameter experiment,the framework is compared with other algorithms for the 500 kV inductor outlet connector and the 220 kV transformer box.The results show that the algorithm proposed in this paper can achieve a good segmentation effect for complex background.Finally,this paper combines the above work with the relative temperature difference method,and designs a set of automatic fault detection system for substation electrical equipment.Through the experiments on the 500 kV inductor outlet connector and the 220 kV transformer box,the fault detection results in this paper have been consistent with the actual,indicating that the system has a good detection effect.