Extração automática de medidas antropométricas a partir de imagens geradas por digitalização a laser e câmeras CCD

• Ano de defesa: 2017
• Autor(a) principal: Pastura, Flávia Cristine Hofstetter
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Civil; UFRJ
• Palavras-chave em Português: Engenharia civil; Antropometria; Digitalização a laser; CNPQ::ENGENHARIAS::ENGENHARIA CIVIL
• Link de acesso: http://hdl.handle.net/11422/9830

This work presents the development of a methodology for anthropometric data collection from human body laser scans. Based on the landmark´s localization, estimation and classification developed methods, it was possible to extract the xyz coordinates of 61 landmarks, 41 1-D measurements, 2-D measurements and 3-D measurements from the scanned image. These methods were tested in 70 digital human models, representatives of both sexes of different biotypes and ethnic origin. The results indicated that the methods are both efficient in the identification of landmarks, with a correct classification rate above 94%, and in the extraction of anthropometric measurements. This makes anthropometric data survey faster and more reliable, as errors inherent in traditional measurement methods such as data variability according to different measurers are eliminated. The contribution of this work is to collect anthropometric measurements quickly, accurately and in a standardized way through automation and dissociation of human interpretation in the collection of anthropometric data. This contribution resulted in a software for automated extraction of anthropometric measurements from Brazilian population scans.