Road junction extraction from high resolution aerial imagery

Road junctions are important components of a road network. However, they are usually not explicitly modelled in existing road extraction approaches. In this research, road junctions are modelled in detail as area objects and an approach is proposed for their automatic extraction through the use of an existing geospatial database. Prior knowledge derived from a topographic geospatial database is used to facilitate the extraction. A new snake-based approach is proposed that makes use of the 'ziplock snake' concept and whose external force is a combination of the gradient vector flow (GVF) force and the balloon force in order to delineate the junction border. Road arm extraction results provide fixed boundary conditions for the proposed snake. The approach was tested using aerial black-and-white Digital Mapping Camera (DMC) ortho-images of 0.1 m ground resolution taken from suburban and rural areas. The results obtained demonstrate the validity of this approach.