基于深度卷积神经网络的货车车号识别研究  

Research on Recognition of Railway Wagon Numbers Based on Deep Convolutional Neural Networks

作　　者：廖健[1] 

机构地区：[1]西南交通大学交通运输与物流学院,成都610031

出　　处：《交通运输工程与信息学报》2016年第4期64-69,80,共7页Journal of Transportation Engineering and Information

摘　　要：准确获取和记录货物列车车号信息是铁路运输系统运行的一项基本任务。传统的货车车号获取办法是由人工抄写记录,存在效率低、容易出现差错等缺点。现有的货车车号自动识别方法需要手工设计图像特征,效率不高。本文利用深度学习技术建立了一种货车车号自动识别方法,首先通过分割边缘密度图像得到字符所在的候选区域,然后通过深度卷积神经网自动学习图像特征,最后利用两个softmax分类器分别完成字符检测和字符识别。识别结果表明该方法可以准确的识别出货车的车号,可以为铁路系统自动化运行和管理提供有力的技术保障。The acquisition of railway wagon numbers is a basic operation task of a railway transport system. The traditional way is to record the numbers manually, which is an error-prone task and has low efficiency. The Existing automatic wagon identification methods require hand-crafted image features, whose efficiency is not high. In this paper, an automatic recognition method for identifying wagon numbers was built via deep learning technology. First, the candidate character region was determined by the edge density image Then, the image features was learnt by the deep convolutional neural networks automatically.Finally, the character detection and recognition were finished by correspondingly. The results shows that the proposed method can accurately. It indicates that the proposed method can provide automatically operate and manage the railway system. two softmax classifiers identify wagons＇ number a technical support to

关 键 词：货车车号识别 深度学习 深度卷积神经网络 

分 类 号：U291[交通运输工程—交通运输规划与管理]