Fault detection in electrical equipment’s images by using optimal features with deep learning classifier

Infrared imaging frameworks have been broadly utilized as a part of the military and civil fields, for example, target recognition, fault diagnosis, fire identification, and medical analysis. Evaluating and monitoring the electrical parts is necessary to analyze the thermal fault at the beginning period. The paper presents the IRT electrical images for diagnosing and classifying the faults by the feature extraction and classification process. At first, IRT segmented switch image (highly temperature zone) is considered, followed by the feature extraction procedure is applied where the images are selected based on the optimal features. The optimal features are accomplished by the inspired optimization algorithm i.e. Opposition based Dragonfly Algorithm (ODA). It chose the best features for the unproblematic classification process. With the intention of classifying the segmented portion as faulty and non-faulty IRT, an approach Deep Neural Network (DNN) is presented. On the basis of the optimal weight attained from learning algorithm, categorize the faulty electrical image easily. The results show that the proposed work accomplishes maximum classification accuracy i.e. 99.99% compared to existing classification approaches.