A Railway Turnout Closeness Monitoring Method Based on Switch Gap Images

Abstract

The railway turnout system is a critical infrastructure that is responsible for steering trains. The closeness state of the turnout is directly related to the safety of the passing trains. To avoid derailment accidents caused by insufficient turnout closeness degrees, it is necessary to monitor the closeness degrees of the turnout. This article introduces a new condition-monitoring strategy that evaluates the turnout closeness degree by measuring the size of the switch gap. To implement this strategy, a detector based on image sensors is designed, and an automatic algorithm is proposed to measure the size of the switch gap based on the images acquired by the detector. The proposed algorithm directly processes the original images of the switch gap and can adaptively extract the region of interest (ROI) and the hysteresis threshold of the Canny operator for each image, which enables workers to avoid using a fixed ROI and hysteresis threshold with different images. Finally, the turnout closeness degree is calculated through a simple conversion of the size of the switch gap, measured by the proposed algorithm. Experiments based on data collected from the actual turnouts of a station show that the proposed turnout closeness monitoring method has high accuracy and robustness, which simplifies the turnout closeness maintenance process and improves its efficiency.