彩色城市栅格地图道路网络自动获取方法  

Automatic Acquiring of Road Network from Color Urban Raster Maps

作　　者：海涛[1,2] 丛爽[1] 鲍远律[1] 

机构地区：[1]中国科学技术大学自动化系,合肥230027 [2]安徽工程大学电气工程学院,安徽芜湖241000

出　　处：《西安交通大学学报》2011年第12期16-21,共6页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(60974092);安徽高校省级自然科学研究项目(KJ2010B023)

摘　　要：为适应栅格地图向矢量地图转化的需求,提出一种从彩色城市栅格地图自动获取道路网络的方法———道路获取对象特征法.由于城市栅格交通地图由道路、区域及噪声3类像素构成,所以新方法包括3个关键步骤:首先利用道路和区域的特征进行图像规范化处理,即所有道路用一种颜色表示,所有区域用另一种颜色表示,图像中其他颜色的像素视为噪声;其次建立4种噪声分类方法来消除规范化图像中的噪声;最后结合规范化图像和道路的连通性,实施道路连接以获取完整的道路网络.与其他方法相比,对象特征法对地图图像的颜色不敏感,可从多类地图图像中自动获取道路网络,并且耗时较少.实验结果表明:对象特征法可完整地提取出地图图像中的道路网络,从整幅合肥市地图图像中提取完整的道路网络仅需32.593s.A novel method for road network automatic acquisition from color urban raster map image,called objects feature method（OFM） of roads acquisition,is proposed to meet the demand of transforming raster maps into vector maps.The proposed method includes three key steps based upon the fact that an urban raster traffic map image is composed of roads,non-road regions and noises.The image is normalized based on the features of the roads and regions,that is,the colors of all roads and regions are replaced respectively with a uniform color.Then the pixels with other colors are regarded as noises,and four noise classification measures are adopted to eliminate the noises in the normalized image.The whole road network is finally constructed by connecting roads from the normalized image and the connectivity of road network.Compared with other methods,OFM is insensitive to colors of geographic objects,and may be applied to some different types of map images.Experiments show that OFM can extract the whole road network from some map images in less time.For example,OFM uses only 32.593 s in extracting the whole road network from the Hefei city color map image.

关 键 词：栅格地图 矢量地图 道路获取 图像规范化 噪声分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]