中文摘要：

随着城市规模的不断扩大,道路标牌的有效管理是智能交通中的一个关键问题.本文在基于本体有机整合海量、多源、异构语义数据的基础上,研究基于国家路标设置规范的路标牌位置和内容有效性自动审核,是对语义技术在智能交通应用中的有益探索.首先,对地缘信息中的道路、兴趣点及路标建立统一模型,采用本体技术把海量数据在语义层面上关联起来.其次,对国家路标设置规范中的规则进行分类,对每一类规则分别构造规则模板及为其构造带参数的SPARQL查询,从而将路标设置指南进行规范形式化表示,在此基础上实现对路标牌的自动检测.最后,在构建的实验数据集上进行了实验,实验结果证明,提出的方法路牌检测准正确率达到85.1%,召回率达到92.1%.

英文摘要：

With the cities ever growing and evolving much faster than before,effectively managing road signs is a major problem for Intelligent Transport System（ITS）.This paper focuses on checking the positioning and contents of road signs in compliance with related road sign regulations（RSRs）based on integrating scalable heterogeneous data.Firstly,road network,POI（point of interest）and road sign are unifiedly modeled,then multi-sources of semantic data are integrated by using mediate ontology.Secondly,rules in RSRs are categorized and parameterized templates and SPARQL queries are constructed for each category,and then validity checking of road signs are based on these formally encoded RSRs.We finally evaluate the proposed method on a real data set.The results show that we achieve the correct rate of 85.1% and the recall rate of 92.1%.A detailed evaluation analysis is presented.