彩色公路地图的道路自动提取

首先采用模板匹配、特征抽取等方法提取城市和道路的标识，这些标识对后面的道路的提取有着重要的作用；然后根据道路的等级，在颜色基础上利用道路的特征分层提取道路图层；最后对道路进行细化，依据城市与道路，各种道路间的关系以及道路的特征建立一系列的判据，检查道路的合理性，并产生相应的策略对道路进行反馈处理，实现道路的全自动提取．实验结果表明了该方法的有效性．

Automatic Road Extraction from Color Road Map

This paper introduces an automatic road extraction method from color road maps. Firstly,the sign symbols,which are very important for later road extraction,are identified from the map by using Labeling and rule-based extraction;Secondly,based on the grades of the roads,road layers are extracted according to color features of the objects on the map,and the extracted roads are thinned.Finally,based on the characteristics of the roads and the relationships between the roads and the sign symbols,this paper also establishes several rules to examine the precision of the roads and mend them if necessary,thus making the road extraction a feedback close looped automatic process.The experiments also demonstrate the method's attectiveness in terms of precision,automation and broad use.