A new approach based on firefly algorithm for vision-based railway overhead inspection system

•A firefly algorithm based fast detection of the pantograph overhead.•Low complexity with accurate detection results.•No required pre-processing method such as segmentation or edge detection.•The ability to distinguish the pantograph overhead in poor illumination conditions.•Detection of arcs by analyzing the overhead region via the Otsu method.

The maintenance of the pantograph–catenary system is a crucial problem in railway inspection systems. The quality of current collection depends on maintaining good contact between a moving pantograph and an overhead contact wire. Problems in the current collection system lead to the damage and disruption of railway traffic. Pantograph arcing occurs between the pantograph and the overhead wire because of the wear of the contact strip. This paper proposes an automatic inspection system for detecting the pantograph and arcs that occur using a firefly optimization algorithm. The proposed method is able to localize the pantograph as a rectangular region and detect the burst of arcing under different illumination and weather conditions simultaneously. The firefly algorithm based Otsu method is used to detect arcs that occur in this region. The proposed approach can serve as an automatic monitoring system and can detect the occurrence of arcing due to loss of contact. The proposed method was applied to four real pantograph videos and efficient results were obtained.