Desenvolvimento de um método para detecção e análise de poeira utilizando processamento de imagens

• Ano de defesa: 2016
• Autor(a) principal: Andrade, Cíntia Ribeiro
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Engenharia Agrícola; UFLA; brasil; Departamento de Engenharia
• Palavras-chave em Português: Speckle laser; Diferenças generalizadas; Momento de inércia; Poeira; Processamento de imagens; Generalized differences; Moment of inertia; Dust; Image processing; Engenharia Agrícola
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/12263

The quantification of suspension dust, one of the most common dispersoids present in the air, has been determined with the objective of fighting human and animal diseases, as well as risks they can bring to the agricultural sector. Despite being precise, the instruments used for the measurement, such as microscopy and sampling pumps, are expensive and of difficult use outside of the laboratories. In order to obtain a method of easy implementation, fast and cheap, this work proposes the detection of dust and other dispersoids by image processing. The images for analysis were captured by a domestic use CCD camera, with and without laser illumination. For image processing, we used the moment of inertia techniques (MI), generalized differences (GD) and coefficient of correlation (ρ). Pilot trials were conducted considering the dust deposited on the surface and the use of the MI technique for analyzing the results. Posteriorly, the same technique was used for analysis, altering the configuration of the experiment in order to maintain the dust in suspension. Finally, still with the dust in suspension, we used the laser illumination and the MI, GD and correlation techniques. The number of replicated, samples and images captured were randomly chosen, seeking the optimization of the results and verifying the limitations of the prototype used in the experiments. The MI values are directly proportionate to the quantity of particulate material in suspension. The GD method allowed the visual evaluation of the increase of suspended material, with the value of coefficient of correlation decreasing with this increase of suspended material. The methods used for optical analysis (MI, GD and ρ) are trustworthy to detect the accumulation of dispersoids suspended in the air.