Development of substation monitoring system (SMS) with improved features for substation automation purposes has become one of the potential areas of research particularly in developing countries like India. The unending SMS automation systems usually incorporate protection of the distribution system from different faults based on monitoring of parameters, such as voltage, current and so forth. But the damaged insulators of the distribution system also affect the system performance very significantly in terms of reduction in voltage and flow of leakage currents. The proposed technique, in contrast to the traditional practice of on-site physical detection, offers a methodology to assess the condition of insulators of an overhead distribution system by means of video surveillance (VS). The algorithm employs modified Hough transform and colour features. Support vector machines (SVM) classifier is trained to classify the presence of a valid insulator in the extracted segments. The most significant contribution of this paper is the development of an efficient algorithm that integrates insulator segmentation from a complex background and thereafter monitoring its condition through VS which makes it an effective tool for real-time purposes in substation automation.
Keywords: video surveillance, support vector machines, SVMs, modified Hough transform, substation monitoring systems, SMS, distribution automation systems, DAS, insulator monitoring, developing countries, India, damaged insulators, overhead distribution systems