Um analisador baseado em visão de máquina para a determinação da concentração de biomassa no processo de fermentação cervejeira

• Ano de defesa: 2023
• Autor(a) principal: Mariz, João Vitor Medeiros
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Química; Programa de Pós-Graduação em Química; UFPB
• Palavras-chave em Português: Fermentação cervejeira; Fermentação alcoólica; Visão de máquina; Processamento de vídeo; Concentração de biomassa; Machine vision; Video processing; alcoholic fermentation; Biomass concentration; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/34035

In this study, we proposed an analytical method with characteristics of speed and sensitivity, developed for the determination of biomass concentration during the beer fermentation process. To achieve this, an analyzer was developed, incorporating lighting control systems and digital film capture. A user-friendly and intuitive control software was developed in the LabView(R) environment, featuring video frame sampling and analysis using machine vision technology, employed for sediment area recognition. Brewer's wort samples were used for calibration and prediction assays. The use of the machine vision algorithm yielded promising results, enabling the construction of a statistical regression model with a coefficient of determination (R2) value of 0.9849, indicating an excellent linear fit. The constructed model was validated through Analysis of Variance (ANOVA), and prediction accuracy was confirmed by paired t-tests and the Joint Elliptical Confidence Region (EJCR) test. The proposed method reduced the analysis time from over 24 h in the reference method to 17 min, eliminating the need for several laborious steps prescribed by the reference method. Thus, the developed instrumentation and method can be considered an interesting alternative for monitoring biomass concentration during the beer production process, as well as paving the way for the implementation of this methodology for other types of fermented beverages. This method stands out for its analysis speed, low implementation cost, and exploration of Industry 4.0 resources, such as intelligent processing, process miniaturization, and machine vision.