钢轨内部伤损检测数据集

钢轨内部伤损会影响行车安全，使用超声波技术进行检测。本数据集来自深圳地铁2号线，世界之窗到侨城北，超声探轮钢轨内部伤损检测系统是基于千兆以太网、由多达 8 个设备机箱堆叠而成的超声波检测系统。每个设备机箱由多达 10 块超声波板卡组成，每块板卡包含 4 个串行硬件通道，每个硬件通道包含 4 个软件通道。板卡之间可以串行、也可以并行。采样使能后，开始发射/接收超声波信号，同时探伤设备将捕获采集数据。FPGA 硬件根据重复频率的设置采集数据。客户机将超声采集数据通过网络传输给计算机系统。数据集包括各个通道A扫、B扫数据，为自定义格式，数据量约80G。

Internal defects in rails pose a significant threat to train operation safety, and ultrasonic testing is the standard technique used for such inspections. This dataset is collected from Shenzhen Metro Line 2, between Window of the World Station and Qiaocheng North Station. The ultrasonic probe wheel-based rail internal defect detection system is a Gigabit Ethernet-enabled ultrasonic testing platform constructed with up to 8 stacked device chassis. Each chassis houses up to 10 ultrasonic acquisition boards. Each board integrates 4 serial hardware channels, and each hardware channel contains 4 software channels. Boards can be configured to operate in either serial or parallel mode. Once the sampling enable function is activated, the system begins transmitting and receiving ultrasonic signals, while the testing equipment captures and acquires the raw data. The FPGA hardware performs data acquisition based on the pre-configured repetition frequency. Client workstations transmit the acquired ultrasonic data to the central computer system via the network. The dataset contains A-scan and B-scan data from all channels, stored in a custom-defined format, with a total data volume of approximately 80 GB.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于钢轨内部伤损检测的超声波检测数据集，来源于深圳地铁2号线，采用基于千兆以太网的多机箱超声波检测系统采集，包含A扫和B扫数据，数据量约80G，旨在支持城市地下基础设施安全监测研究。数据集属于国家重点研发计划项目，专注于超声学在轨道交通领域的应用，为钢轨内部损伤的检测和分析提供基础数据。