Automatic Extraction of Road Networks based on Normalized cuts and Mean Shift method for high resolution Satellite Imagery

To evaluate the speed of growth of an urban area road is one of the fast information updating element during urban development. Road information extraction based on high resolution satellite images play an important role because roads affect city land usage. In this paper, two approaches for road network extraction for an urban are proposed. Most research in road extraction begins with an original image. It is difficult and computationally expensive to extract roads due to presences of other road-like features with straight edges. In the proposed method firstly the image is preprocessed to improve the tolerance by reducing the noise (the buildings, parking lots, vegetation regions and other open spaces). Secondly roads are extracted based on Normalized cuts method and Mean Shift Method. Finally the accuracy for the road extracted images is evaluated based on quality measures. The experimental results show that these approaches are efficient in extracting road segments in urban region from high resolution satellite images. KeywordsRoad extraction; Filtering; Mean shift; Normalized cut; performance evaluation