Gap Detection of Switch Machines in Complex Environment Based on Object Detection and Image Processing

A turnout, a device to guide tracks, is critical to the safety of high-speed railways. Detecting gaps in switch machines can monitor a turnout's working performance. Existing gap-detection systems, however, can barely perform at high accuracy and with a low false alarm rate for a long time due to the complex operating conditions of switch machines. This study proposes an approach combining YOLO-based object detection architecture with image processing algorithms, of which YOLO is a deep learning network for object detection. First, YOLO detects target areas in gap images, and then image-processing algorithms identify gaps and calculate gap sizes. This approach targets various types of switch machines and particularly complicated situations. Experiments on gap images of S700K switch machines demonstrate that the accuracy of detecting gaps is 100%, and the accuracy of calculating gap sizes is higher than 99%. Additionally, the proposed approach can exhibit the same high performance on complex images, like overexposed and tilted ones.