Desenvolvimento e avaliação de um protótipo classificador de tábuas usando técnicas de visão artificial

• Ano de defesa: 2007
• Autor(a) principal: Gomes, José Marcelo
• Orientador(a): Pinto, Francisco de Assis de Carvalho
• Banca de defesa: Fernandes, Haroldo Carlos , Santos, Nerilson Terra , Khoury Junior, Joseph Kalil
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Doutorado em Engenharia Agrícola
• Departamento: Construções rurais e ambiência; Energia na agricultura; Mecanização agrícola; Processamento de produ
• País: BR
• Palavras-chave em Português: Processamento de imagens; Classificação de tábuas; Inspeção automatizada
• Palavras-chave em Inglês: Image processing; Lumber classification; Automated inspection
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA::MAQUINAS E IMPLEMENTOS AGRICOLAS
• Link de acesso: http://locus.ufv.br/handle/123456789/764

Much effort has been expended by researchers to apply automation to productive processes. The sawn wood industry has a promising place in this activity, because lumber is a raw material for products with increasing demand, such as furniture, which requires high quality pieces. The quality of the lumber is determined by the defects it presents and some characteristics correlated to these defects such as: size, position, amount and type. The goal of the present research was to develop and evaluate a prototype to classify eucalyptus lumber using digital images. A classifier was developed and tested to recognize defects in eucalyptus lumbers taking in consideration the histogram values of the percentiles of pixels, morphologic information and position of the defects. Image processing and analysis for lumber classification was implemented in the classification algorithm. The following step was the construction and testing of a prototype for the classification of lumber using a machine vision system. It is built in a conveyor belt where the lumbers are inserted and conducted for image acquisition. The prototype can use both, the Brazilian standard (ABNT) and a commercial rule for lumber classification. The process can be followed in the microcomputer screen that shows the lumber images with its final grade. The rules were simplified to make possible its application. The overall accuracy in the classification process was 64 and 81% percent using the ABNT and commercial rules, respectively. The productivity of the developed prototype was 7.9 m3 h-1 in eucalyptus lumber grading.