A Robust Anomaly Detection in Pantograph-Catenary System Based on Mean-Shift Tracking and Foreground Detection

Abstract

This study presents a robust condition monitoring and anomaly detection for pantograph-catenary system. The pantograph overhead system is monitored by using a digital camera. A general framework for anomaly detection for pantograph-catenary system consists of two key components. The first component is based on mean-shift tracking of contact wire. Therefore, the contact point between pantograph and catenary can be continuously monitored and anomaly contact to some points will be detected. The second component uses Gaussian mixture model (GMM) for foreground detection. When the foreground of the current frame has been detected, the mean-shift tracking and GMM combines trajectory-based and region-based information for detection any anomaly in pantograph-catenary interaction. The experimental results show that proposed method is useful to detect burst of arcing, and irregular positioning of contact line.