Road extraction in urban areas from high resolution satellite images helps in creating a database of a city and in cartography. The extraction results are intended to be used for updating a road database. The high dimensionality of aerial and satellite imagery presents a challenge to the human analysis based on the traditional classification algorithms using statistical assumptions. Artificial Neural Networks (ANNs) on the other hand may represent a valuable alternative approach for land cover mapping for such highly dimensional imagery. The urban areas contain roads of different shapes, sizes and lengths. In this paper, the extraction algorithm performs edge detection, morphological reconstruction, feature extraction and classification. The road features are classified using ANNs.