Extração automática de cruzamentos simples e complexos de rodovias em imagens digitais de alta resolução

• Ano de defesa: 2012
• Autor(a) principal: Zanin, Rodrigo Bruno [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Cartografia; Rodovias; Imagens digitais; Topografia; Espectroscopia de passagem de nivel; Cartography
• Link de acesso: http://hdl.handle.net/11449/108616

The first works related to road extraction in digital images already have over thirty years. However, research related to the extraction of roads and consequently the road network are still in expanding, integrating methods from the fields of Digital Photogrammetry, Digital Image Processing and Image Analysis. This is result of the great importance that the process of extraction of road network present in numerous applications, such as updates to Geographic Information Systems (GIS), road planning and fleet monitoring. Among various challenges related to research on extraction of road network, certainly one can include the extraction of road junctions. Considering the complexity and diversity of road crossings, there is little information, a priori, that can be used in creating a model for extraction of road junctions. In this context, this paper proposes a method for extracting road junctions in highresolution digital images, classified as simple and complex junctions. The proposed method uses some traditional processes in the area of Image Analysis, dividing itself into three sequential steps to be applied in regions containing road junctions...