Improved image processing approaches to detect electrical system faults by extracting and analyzing features from infrared thermography images

Abstract

A major cause of failure in electrical equipment is the effect of heating due to circuit malfunction. Abnormalities within the instrumentality can occur once their internal temperatures exceed their limits. The common issues concerning thermal anomalies in electrical installations are loose or poor connections, unbalanced loads, short circuits, overloading and cracks or defects within the instrumentality body. Infrared thermography (IRT) is a well-known effective tool for monitoring the condition of equipment. With the thermal images, inspectors can analyse the temperature variations of thermal objects to look for defective parts. In addition, the effective maintenance of equipment in most cases would require the presence of well-experienced and trained personnel in order to analyse the infrared images and perform an accurate prediction. However, this might not be an optimum solution due to a shortage of staff. Hence, by feeding the images to an intelligent system, an accurate diagnosis can be obtained and read even in the absence of experts. Rapid development in image processing techniques and the integration of artificial intelligence has provided advantages in monitoring and diagnosing problems with the electrical equipment. The objective of the research focuses on developing innovative image processing algorithms to localize and classify the faults in electrical equipment by examining the thermal images of the equipment and thereby by monitoring its healthiness. The procedure utilized for fault discovery in electric equipment examined by preprocessing, segmentation, and feature extraction and classification strategies with powerful optimization algorithms. newline