摘要
为实现地铁车辆基地内车辆编号的自动识别与采集,综合深度学习、目标检测、机器视觉及光学字符识别技术,利用车辆标志与编号相对位置不变的特点,研发一种适用于地铁车辆的车号识别系统。经过现场安装调试测试,车号识别准确率及可靠性均满足现场使用要求。
In order to realize the automatic identification and collection of vehicle numbers in subway rail yards,a vehicle number recognition system suitable for rail yards is developed in this paper.This system integrates deep learning,target detection,machine vision and optical character recognition technologies,and takes advantage of the invariable relative position of vehicle markings and numbers.After installation,debugging and testing,the accuracy and reliability of vehicle number recognition system meet the requirements of on-site use.
出处
《现代城市轨道交通》
2022年第5期20-23,共4页
Modern Urban Transit
基金
陕西省重点研发计划项目(2020GY-182)。
关键词
地铁
车辆基地
车号识别
深度学习
目标检测
metro
rail yard
vehicle number recognition
deep learning
target detection
