轨道交通病害多模态检测数据集

本数据集数据采集分为内部和外部两种方式。内部采集主要依靠公司自身在轨道和隧道中部署的专业设备，如多模态传感器，包括激光雷达、高清摄像头、振动传感器等，按照预设的检测路径和参数，实时采集轨道和隧道的多模态数据，并通过内部网络传输至数据中心。外部采集则涉及与其他科研机构或企业合作，获取其在类似场景下采集的相关数据，用于丰富数据集，提升模型的泛化能力和检测精度。This dataset is gathered through two distinct collection modes: internal and external. For internal collection, professional devices deployed by the company itself in rails and tunnels are utilized, including multimodal sensors such as LiDAR, high-definition (HD) cameras, and vibration sensors. Multimodal data of rails and tunnels are collected in real time in accordance with preset detection routes and parameters, and then transmitted to the data center via the internal network. For external collection, the dataset is enriched by cooperating with other research institutions or enterprises to acquire relevant data collected by them in similar scenarios, which helps improve the generalization ability and detection accuracy of the model.

搜集汇总

背景与挑战

背景概述

轨道交通病害多模态检测数据集由武汉纵横天地空间信息技术有限公司提供，包含1000条多模态数据，用于轨道和隧道的病害检测，更新频次为每半年。数据集通过内部和外部采集方式获取，应用场景包括轨道健康监测、隧道病害检测和设备故障诊断。