基于Faster R-CNN的铁路超限货物自动检测系统  

Automatic detection system of railway out of gauge goods based on faster r-cnn

作　　者：庞婷婷[1] 郝菊香 PANG Tingting;HAO Juxiang(Xi’an traffic Engineering institute,Xi’an 710300,China)

机构地区：[1]西安交通工程学院,西安710300

出　　处：《自动化与仪器仪表》2021年第8期72-76,共5页Automation & Instrumentation

基　　金：西安交通工程学院中青年基金项目:铁路超限货物装载检测技术研究(No.21KY-10)。

摘　　要：目前铁路超限货物自动检测系统,未曾预处理设备采集到的检测图像,导致系统检测铁路超限货物时,存在检测误差大、精度低、实际应用效果差等问题,为此提出基于Faster R-CNN的铁路超限货物自动检测系统。硬件设计方面,设计系统硬件结构。软件设计方面,预处理检测图像,引入Faster R-CNN,自动监测铁路超限货物。实验结果表明:设计三组系统的对比实验,此次设计的检测系统,具有较小的检测误差、较高的检测精度、较优的实际应用效果。at present,the automatic detection system of railway out of gauge goods has not preprocessed the detection image collected by the equipment,which leads to the problems of large detection error,low accuracy and poor practical application effect when the system detects railway out of gauge goods.Therefore,the automatic detection system of railway out of gauge goods based on fast r-cnn is proposed.In terms of hardware design,the hardware structure of the system is designed.In terms of software design,the image is preprocessed and fast r-cnn is introduced to automatically monitor the out of gauge goods.The experimental results show that:the design of three groups of system contrast experiments,the design of the detection system,with smaller detection error,higher detection accuracy,better practical application effect.

关 键 词：Faster R-CNN 铁路 超限货物 货物自动检测 检测系统 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] U294.6[自动化与计算机技术—控制科学与工程]