高铁应用识别测试数据集

数据来源：数据采集具体位置在轨旁7m处，首先依次连接好数据采集系统：200w像素高清晰度超高速脉冲相机连接电源线与数据传输线，数据传输线另一端连接数据传输主机；数据传输主机连接电源线、外界显示器HDMI传输线、无线键鼠以及数据传输线；整体通过电小二与插排对数据传输主机、相机以及外接显示器供电，打开主机，打开数据采集软件，将相机焦距与光圈调整到合适位置，等待高铁经过时进行数据保存，利用笔记本电脑进行原始脉冲数据的重构。 采集地点：京广高铁线路长沙-武汉段K1565+790。 采集时间：2024年11月10日。 数据集内容：测量超高速脉冲视觉高铁关键部件实时监测软件对高铁关键零部件的识别率，包括测试集原始脉冲数据、重构视频数据、训练数据集、三方测试数据、测试结果数据、测试大纲及三方测试报告。数据量共238GB。Data Source: The data was collected at 7 meters beside the railway track. First, set up the data acquisition system in sequence: connect the power cable and data transmission cable to the 2-megapixel high-definition ultra-high-speed pulsed camera, then connect the other end of the data transmission cable to the data transmission host. Connect the power cable, HDMI cable for external display, wireless keyboard and mouse, and an additional data transmission cable to the data transmission host. Power the data transmission host, camera and external display via a DianXiaoer portable power station and a power strip. Turn on the host, launch the data acquisition software, adjust the camera's focal length and aperture to appropriate positions, save data when a high-speed train passes by, and reconstruct the raw pulsed data using a laptop. Collection Location: K1565+790, Changsha-Wuhan Section of Beijing-Guangzhou High-Speed Railway Line. Collection Time: November 10, 2024. Dataset Content: This dataset is used to measure the recognition accuracy of the ultra-high-speed pulsed vision-based real-time monitoring software for key components of high-speed trains. It includes raw pulsed data in the test set, reconstructed video data, training dataset, third-party test data, test result data, test outline and third-party test report. The total data volume of the dataset is 238 GB.

搜集汇总

背景与挑战

背景概述

该数据集用于测试超高速脉冲视觉软件对高铁关键零部件的识别率，包含原始脉冲数据、重构视频、训练集、测试结果等，数据量达238GB。数据于2024年11月10日在京广高铁长沙-武汉段采集，使用高速脉冲相机在轨旁7米处获取。