地铁车辆障碍物检测技术应用研究

障碍物检测是确保地铁车辆运行安全的一个重要措施,但是目前地铁运行障碍物检测存在场站环境下单一传感器检测速度慢、准确性差、范围小的问题,为实现对其的有效弥补,在可见光与近红外结合的双目视觉技术迅速发展的支持下,基于红外成像系统与可见光成像系统,设计地铁障碍物检测系统;基于对系统功能需求与运行平台的分析,明确系统功能模块及其硬件构成;由于红外与可见光图像传感器的成像机理不同,拍摄的图像也存在差异,利用BEEMD算法对图像进行融合之后,执行YOLO障碍物类别检测任务;最后分析系统应用,结果表明可以实现列车运行中前方列车80～500 m安全限界范围内的障碍物识别。

Application of subway obstacle detection technology based on image recognition technology

Obstacle detection is an important measure to ensure the safety of subway vehicles. However, at present, the detection of obstacles in subway operation has the problems of slow detection speed, poor accuracy and small range by a single sensor in the station environment. In order to make up for it effectively, with the support of the rapid development of binocular vision technology combining visible light and near-infrared, a subway obstacle detection system is designed based on infrared imaging system and visible light imaging system. Based on the analysis of the system functional requirements and operating platform, the system functional modules and hardware composition are defined. Due to the different imaging mechanism of infrared and visible image sensors, the images taken are also different. Finally, the application of the system is analyzed. The results show that the obstacle recognition in front of the train within 80-500m safety limit can be realized.