Thermal imaging for qualitative-based measurements of thermal anomalies in electrical components

Abstract

This paper proposes a method of classifying the reliability of electrical equipment by analyzing their thermal images. In order to automatically analyze the thermal image, a top-down approach of image processing is used. First the distinctive feature points of the target equipment are identified. The maximally stable extremal region (MSER) algorithm is used to detect the feature points and regions of interest. Feature descriptors for each detected point are calculated and similar features are matched together by utilizing the Euclidean distance to find similar equipment within the image. These are then grouped together before proceeding with the segmentation process. The condition of the electrical equipment is evaluated by finding their real temperature values. Classification of the thermal faults within the electrical equipment is done by using qualitative-based measurements. The results indicate that this technique produces about 60% correct classifications, which is according to the recommended standards.