Intelligent detection of electrical equipment faults in the overhead substations based machine vision

Abstract

This paper introduces a method for the intelligent detection of electrical equipment faults with thermo vision technology, multi-class support vector machine (MSVM) as a classifier and Pseudo Zernike moment as image feature. The aim of this paper is to detect the electrical equipment faults by making use of the moment method and statistical features of thermo images. The classifier effectiveness and accurateness depends on the moment order that was used. By attention to the commonly occurring faults in the substations of distribution networks, four major faults occurring in overhead substations have been chosen. Simulation results are carried out on practical databases of real images of the distribution networks of North West of Tehran.