Desenvolvimento de um multianalisador automático baseado em visão de máquina e processamento de vídeo para determinação sequencial do teor de sólidos suspensos totais e sedimentáveis em águas residuárias

• Ano de defesa: 2021
• Autor(a) principal: Ramos, Railson de Oliveira
• Tipo de documento: Tese
• 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: Visão de máquina; Processamento de vídeo; Instrumentação; Análise de águas residuais; Sólidos suspensos; Sólidos sedimentáveis; Machine vision; Video processing; Instrumentation; Wastewater analysis; Suspended solids; Settleable solids; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/21564

The monitoring of suspended solids (SST) and sedimentable solids (SS) in water and effluents is of great importance for process control and quality management in environmental systems. Nevertheless, the standard methods for determining these parameters comprise a laborious analytical gait, with a high response time and costly instrumentation. In view of this, in the present study, two analytical methods with characteristics of speed and sensitivity were proposed, developed for sequential determination of the concentration of total suspended solids (SST) and sedimentable solids (SS). Therefore, a multianalyzer equipped with an automatic sampler, lighting control, laser radiation source and digital film capture system was developed. In addition, a control software was also developed, which included a video frequency sampling system, resources for colorimetric analysis by digital images and machine vision analysis. RGB, HSL, HSV and HSI color spaces were evaluated in the video pre-processing phase. Wastewater samples were used in the calibration and prediction tests. In determining the SS concentration, the machine vision system was used to recognize and determine the area formed by the sediments. In determining the concentration of SST, the machine vision system was used to recognize and determine the area of light scattered by the suspended particles. The machine vision algorithm coupled to the red color plane (derived from color histograms in the Red-Green-Blue (RGB) system) showed the best results with R2 of 0.997 and 0.988 and RMSEP of 3.188 mg L-1 and 0.300 mL L-1 for the determination of SST and SS, respectively. The constructed models were validated by Analysis of Variance (ANOVA) and the accuracy of the predictions confirmed by the joint confidence elliptical region test (EJCR). In relation to the reference method, the proposed method reduced the sample volume from 3.5 L to just 15 mL and the analysis time from 12 hours to 24 seconds per sample. Therefore, the methods that were obtained can be considered an important alternative for the determination of SST and SS in wastewater, as an automatic, fast and low-cost way, in line with the principles of Green Chemistry and exploiting Industry 4.0 resources such as intelligent processing , miniaturization and machine vision.