Road Extraction from Color Raster Urban Triffic Map

In recent years, with the aid of computer technology and geographical information system (GIS), which has the ability to handle a large amount of geographical data in it. In particular, substantial efforts are devoted to the conversion of paper-based map into computer readable from which is suitable for input into GIS. This paper puts forward a road extraction algorithm, which includes three key steps. Firstly, the map image is normalized taking advantage of the existing normalizing algorithms in some literatures. Secondly, based on the features of the noises of the normalizing map such as the expand feature of the noises and so on, some classification measures of the noises are built and adopted to recognize and classify the noises. Finally, the whole procedure of the road extraction is given, and some experiments are applied to validate the feasibility and validity of the algorithm. It is shown that the algorithm is effective in various cases successfully.