CLASSIFICAÇÃO DE CERVEJAS POR ANÁLISE DE IMAGENS E PCA

• Autor(a) principal: SILVA, Thiago César de Oliveira
• Data de Publicação: 2008
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• dARK ID: ark:/38995/0013000006gx9
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tde/1035
• topic: PCA; Análise de Imagens; Cerveja; 1. Cerveja - Caracteristica 2. Cerveja - Controle de qualidade 3. Análise de imagens 4. Análise de componentes principais; PCA; image analysis; lager beer; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA::FISICO-QUIMICA::QUIMICA TEORICA

In this work a simple and new approach for lager beer pattern recognition based on image analysis was accomplished using Principal Component Analysis (PCA). Digital color images obtained from a scanner were decomposed into a binary form with the help of Scilab and SIP softwares. The PCA data matrix consisted of RGB color histograms for each sample. 350 mL beer cans were chosen at random on the shelf of local supermarkets in an overall of ninety cans from ten different brands. 200 mL samples were immersed in an ultrasound bath for a period of time ranging from twenty to thirty minutes before the digital image processing. The resulting methodology can be a simple alternative to the analytical quality control of lager beers.