Application of Time Series Algorithms in Infrared Image Segmentation for Fault Detection and Defect Identification of Substation Equipment

To improve the accuracy of fault defect identification, image clarity, and detail integrity, an infrared image segmentation method for fault defects of substation equipment based on time series algorithm is proposed. A framework for infrared remote vision and fault defect identification of substation equipment is designed. Suspicious fault defect images are screened and marked through a sliding window combined with time series. After noise removal using wavelet threshold transformation and median denoising method, the improved watershed algorithm is used to complete image target segmentation. Experimental results show that after denoising, the image clarity and contrast are significantly improved, and the detail integrity of the target equipment after segmentation is high. Different equipment fault defects can be identified.