基于改进YOLOv5的道岔位移检测方法研究

Research on turnout displacement detection method based on improved YOLOv5

摘要:

针对地铁系统，设计了一种创新的道岔位移检测方法，该方法融合了优化的YOLOv5目标检测算法与二维码位置识别技术；通过精准识别道岔上的二维码位置变化，实现位移的实时检测与预警；通过在轨道叉尖上粘贴二维码标靶，利用视觉传感器实时检测叉尖位移；系统使用YOLOv5s模型进行目标检测，并引入CBAM注意力机制和DIoU损失函数，提高检测精度和效率；为了提高道岔位移检测的实时性，在异常情况突发下及时报警，在网络中添加了ShuffleNet V2模块；实验结果表明，改进后的YOLOv5模型在不同光照条件下对二维码标靶的检测性能优异，实时性满足使用需求，为地铁轨道的健康检测提供了可靠的数据支持；此方案旨在克服传统检测方法中的局限性，确保列车运行的更加平稳与安全；

Abstract:

For the subway system, An innovative turnout displacement detection method was designed , which integrated the optimized YOLOv5 object detection algorithm and QR code position recognition technology. By accurately identifying the position change of the QR code on the turnout, real-time detection and early warning of displacement can be realized; By pasting the QR code target on the track fork tip, the fork tip displacement is detected in real time by using a vision sensor; The YOLOv5s model is used for object detection, and the CBAM attention mechanism and DIoU loss function are introduced to improve the detection accuracy and efficiency. In order to improve the real -time nature of the road fork shift detection, the police reported in time under abnormal situations, we added the ShuffNet V2 module to the network. Experimental results show that the improved YOLOv5 model has excellent detection performance on QR code targets under different lighting conditions and meets the needs of use in real time, which provides reliable data support for the health detection of subway tracks. This solution aims to overcome the limitations of traditional detection methods and ensure a smoother and safer train operation.