Segmentação de cristais de clínquer em imagens microscópicas via redes neurais convolucionais

• Ano de defesa: 2024
• Autor(a) principal: Rodrigues, Renan Vinicius
• Orientador(a): Zuanetti, Daiane Aparecida
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa Interinstitucional de Pós-Graduação em Estatística - PIPGEs
• Palavras-chave em Português: Automação de processo; Aprendizado por transferência; Desempenho preditivo; Filtros de pré-processamento
• Palavras-chave em Inglês: Transfer learning; Process automation; Predictive performance; Preprocessing filters
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::ESTATISTICA::ANALISE DE DADOS
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/21166

Portland cement is currently carried out by trained professionals who analyze the crystals present in the microstructure of the clinker (an input produced in the cement manufacturing process and which gives it its main characteristics). Among these crystals, the one that most affects the final product is Alita (C3S). Because of this, building an automatic process for segmenting and classifying C3S in microscopic images of clinker can bring savings and efficiency in cement manufacturing. This work, therefore, seeks, through convolutional neural networks and image pre-processing filters, to carry out this segmentation so that the automation of the process is viable, increasing the quality of the product. A description of neural networks and their extensions is provided, as well as a brief review of the most common image preprocessing filters. Subsequently, several neural network models are fitted and compared in the analysis of clinker images.