Abstract: A novel method for road extraction from color raster highway traffic maps, called neighborhood method, is proposed to meet the demand of transforming raster maps into vector maps.Firstly, a feature-matching algorithm is applied to recognize and extract automatically graphic symbols of all county towns.Secondly, a measure is put forward to extract the candidate road from each neighborhood of the extracted graphic symbols, and then some features of roads and noise are utilized to eliminate all noise and misrecognition of roads.Finally, a number of experiments evaluate qualitatively and quantitatively the effectiveness and superiority of the proposed method.Compared with other methods, the proposed method is high-efficiency and insensitive to colors of geographic objects, and may be applied to some different types of color raster highway traffic map images.