URBAN ROAD EXTRACTION FROM HIGH RESOLUTION REMOTE SENSING IMAGERY USING MEAN SHIFT-AIDED ACTIVE CONTOUR MODEL

摘要

Road extraction from High Resolution remote Sensing Imagery (HRSI) for urban planning and Geographic Information System (GIS) databases update is an ongoing research topic in the field of remote sensing imagery processing. In lower resolution satellite images, roads are normally characterized by linear features, while in high ones roads are mainly areas with regular width or shapes, therefore, traditional linear feature extraction methods cannot be used in HRSI indiscriminately. In this paper, a novel mean shift-aided active contour model combining two image processing methods is proposed of which one is mean shift image segmentation and the other is region-based active contour model for road extraction. The results with the high resolution aerial orthoimages covering some parts of the city of Las Vegas, Nevada, USA proved the effectiveness of the proposed approach.