Orientação semi-automática de uma sequência de pares de imagens frontais por fototriangulação a partir de fotocoordenadas extraídas pelo SIFT

• Ano de defesa: 2014
• Autor(a) principal: Lemes Neto, Maurício Correia [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Cartografia; Fotogrametria; Imagens digitais; Sistemas de navegação global por satélite; Sistemas de navegação inercial; Mapas - Escalas; Cartography
• Link de acesso: http://hdl.handle.net/11449/127591; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/20-08-2015/000843718.pdf

Systems Terrestrial mobile mapping systems integrate inertial navigation and positioning sensors such as INS (Inertial Navigation System), GNSS (Global Navigation Satellite System), digital cameras and methodology can create georeferenced imagens and coordinates from any attribute of observable objects. However, problems such as loss of GNSS signal and the errors accumulation's by the continuous use of the INS can directly influence the orientation and positioning of digital images obtained by photogrammetric surveying. With the goal of adding more resources and alleviate the problems caused by the limitations of the sensors this thesis proposes a semi-automatic extraction solution of homologous points in stereo image pairs to perform photo-triangulation based in observations obtained only of the images and local reference arbitrated. The technique of extracting points in digital images used in this study is the SIFT (Scale Invariant Feature Transform) and the correspondence among them is taken from a search for similarity descriptors of each key point whose metric analyzed is the smallest Euclidean distance. There may be false matches proposed by David Lowe method compares the shortest distance with the second best distance selecting only the corresponding points that are near a threshold previously established...