列车车底现场螺栓巡检数据

在地铁运营中，列车车底螺栓的状态直接关系到列车运行的平稳性与安全性。通过车辆智能巡检机器人采集列车车底的巡检图像，并识别图像中螺栓，分析列车车底设备螺栓的状态，实现列车车底设备螺栓的预警与报警，提升列车车底设备螺栓状态的自动化、智能化监测水平，大大减少了人工巡检的工作量，有效降低了列车运行过程中因螺栓故障引发的安全事故风险，为列车的安全、稳定运营提供了有力保障。通过实时上传采集列车车底的巡检图像，分析列车车底设备螺栓的状态，实现列车车底设备螺栓的预警与报警，具体过程：
（1）数据采集：机器人行驶至指定导航点位，并确认列车是否位于指定的轴区域内，控制执行机械臂至预设位置后，并结合相机预设参数（如倍率、对比度、饱和度）及补光灯亮度参数进行拍摄成像；
（2）数据处理：拍摄图像传入算法识别模块，先通过深度学习YOLOv8目标检测模型进行螺栓的检出，再通过Unet++分割网络对螺栓防松线进行提取算法操作，通过防松线的判断算法（是否错位）模块进行分析处理，并结合算法模版图比对降低识别误报率，综合判断螺栓设备状态是否正常；
（3）将判断分析后的螺栓设备状态值上传至界面进行结果展示，设备状态显示：松动，预警类型为1。设备状态显示：正常，预警类型为0。实现螺栓设备的预警与报警，便于现场人员快速判别设备的状态是否异常；
（4）同时可根据采集所得的数据按天、月、年等时间维度进行趋势分析，形成安全报表数据。

In subway operations, the condition of train undercarriage bolts directly affects the stability and safety of train operations. Intelligent inspection robots for rail vehicles collect undercarriage inspection images, identify bolts in the images, analyze the condition of undercarriage equipment bolts, realize early warning and alarm for undercarriage equipment bolts, improve the automated and intelligent monitoring level of undercarriage equipment bolt conditions, greatly reduce the workload of manual inspection, effectively lower the risk of safety accidents caused by bolt faults during train operation, and provide a strong guarantee for the safe and stable operation of trains. By real-time uploading of collected undercarriage inspection images to analyze the condition of undercarriage equipment bolts and realize early warning and alarm, the specific workflow is as follows:
(1) Data Collection: The robot travels to the designated navigation point, confirms whether the train is within the specified axle area, moves the executive robotic arm to the preset position, and captures images with preset camera parameters (e.g., magnification, contrast, saturation) and fill light brightness parameters;
(2) Data Processing: The captured images are sent to the algorithm recognition module. First, the deep learning YOLOv8 object detection model is used to detect bolts, then the Unet++ segmentation network is applied to extract the anti-loosening wires of the bolts. Analysis is conducted via the anti-loosening wire judgment algorithm (misalignment detection) module, and the false recognition rate is reduced by comparing with algorithm template maps, so as to comprehensively judge whether the bolt equipment is in normal condition;
(3) The judged and analyzed bolt equipment status values are uploaded to the interface for result display. The equipment status is displayed as "Loose" with warning type 1, and "Normal" with warning type 0. This realizes early warning and alarm for bolt equipment, enabling on-site personnel to quickly judge whether the equipment status is abnormal;
(4) Additionally, trend analysis can be conducted on the collected data according to time dimensions such as day, month and year, to generate safety report data.

搜集汇总

背景与挑战

背景概述

该数据集为列车车底现场螺栓巡检数据，包含501条xlsx格式记录，每日更新。通过智能巡检机器人采集图像，利用YOLOv8和Unet++算法分析螺栓状态，实现自动化预警，提升地铁运营安全性。