Robust real-time contact-point detecting and tracking by a horizontal-vertical enhancement method

The pantograph-catenary interaction is a safety indicator of electric multiple unit (EMU) trains. It is important for EMU trains to monitor the contact-point of pantograph and catenary in real time. In this paper, a highly efficient contactpoint- tracking approach, called horizontal-vertical enhancement and tracking method (HEATM), is proposed to enhance infrared images which are used to detect and track contact-point. The HEATM includes the following three key components. Firstly, the HEATM separates the input infrared image into the horizontal images (HIs) layer and vertical images (VIs) layer. Secondly, the contact-wire tracking model is updated by the points from HIs while the pantographtracking model is updated by the points from VIs. Finally, the key contact-points are positioned by the tracking model and continuously analyzed to obtain robust tracking signals. The quantitative and qualitative results validate the effectiveness of the proposed scheme. Moreover, our method presents very robust and efficient for contact-point detecting and tracking at a speed of 70 fps and 97.8% average accuracy in two datasets (12000 frames), which is very beneficial for its extensive application.