There exist two options for digital cameras that can capture the near-infrared (NIR) band. Conventional red-green-blue (RGB, visible bands) cameras with a single sensor provide NIR band visibility based on the removal of the internal NIR-blocking filter. Alternatively, multisensor cameras exist that have a specific sensor for each band. The modified RGB cameras are of a lower price. In this context, the objective of this study was to compare the performance of a modified RGB camera with that of a multisensor camera for obtaining the normalized difference vegetation index (NDVI) in an area with coffee cultivations. A multispectral camera with five sensors and another camera with only one sensor were used. The NDVI of the coffee field was also measured using the GreenSeeker handheld NDVI sensor manufactured by Trimble. The images were calibrated radiometrically based on the targets in shades of gray made of napa, and the NDVI was calculated after image calibration. The calibration curves showed a high coefficient of determination. The NDVI value obtained with the calibrated images from the cameras showed a significant correlation with the values obtained by the GreenSeeker NDVI sensor, making it possible to obtain the variability pattern of the vegetation index. However, the NDVI obtained using the multisensor camera was closer to the NDVI obtained by the GreenSeeker NDVI sensor.

The selection of coffee beans plays a key role in the product's final quality. After processing, coffee beans are classified according to their quantity of defects. Traditionally this classification is performed manually, which makes the process laborious and time-consuming. This problem can be solved with digital image processing techniques since defective grains have unique visual characteristics. Considering the difficulty of manual classification of the defects, this study aimed to elaborate a Bayesian classifier algorithm to identify these defects in benefited coffee beans, based on its shape and color. To do so, 630 grains of arabica coffee were used, composing eight images in total. The algorithm aimed to classify four classes, which were: regular beans, normal broken beans, black beans, and black broken beans. In order to evaluate the accuracy of the classifier algorithm, it was calculated the global accuracy and the Kappa coefficient, which allows inferring if the classifier is better than a random classification. It was concluded that the developed algorithm presented a global accuracy of 76% and kappa equals to 0.6. Also, the proposed methodology showed great potential for application in the quality evaluation of other products, whose shape and spectral parameters are relevant in evaluating its quality.

Níveis de ruído emitidos por tratores agrícolas com e sem acionamento de implementos Os níveis de ruído emitidos pelas máquinas agrícolas podem ser prejudiciais à saúde do trabalhador e, por vezes, negligenciados no meio rural. Este trabalho teve como objetivo mensurar os níveis de ruído emitidos por três tratores agrícolas de diferentes potências acionando dois tipos de implementos para avaliar se havia riscos à saúde dos operadores e trabalhadores de apoio ao redor da operação. Os dados de ruído foram coletados por decibelímetro em dez pontos de cada lado da máquina (frente, trás, esquerda e direita) para um total de 40 pontos ao redor de cada trator. Três tratores diferentes foram avaliados em diferentes cenários: sem implemento, acoplando e acionando uma esparramadora de corretivo e acoplando e acionando uma enxada giratória. O trator foi estacionado no centro da mistura com uma rotação do motor que garantiu 540 rotações por minuto (RPM). para a tomada de potência (TDP) durante toda a coleta de dados. Os dados foram analisados ​​por meio de gráficos, regressão linear e análise de agrupamento hierárquico. Os resultados indicaram que os níveis de pressão sonora em todas as situações estudadas ultrapassam os limites da norma regulamentadora (NR) 15, tornando os protetores auditivos indispensáveis ​​durante a jornada de trabalho. Palavras-chave: ergonomia; conforto acústico; mecanização agrícola; nível de pressão sonora.   ABSTRACT: The noise levels emitted by agricultural machines can be harmful to the worker's health, and it is sometimes neglected in rural areas. This work aimed to measure the noise level emitted by three agricultural tractors of different power activating two types of implements to assess whether there were risks to operators' health and the supporter workers around the tractor. The noise level data were collected using a decibel meter in ten points on each machine side (forward, rear, left, and right) for a total of 40 points around each tractor. Three different tractors were evaluated in different scenarios: without any implement, coupling and activating a spreader, and coupling and activating a rotary hoe. The tractor was parked at the centre of the mash with an engine speed that ensured 540 rotation per minute (RPM). to the power take-off (PTO) during the entire data collection. The data were analyzed by charts, linear regression, and hierarchical clustering analysis. The results indicated that the sound pressure levels in all of the studied situation exceed the standard's limits regulatory standard 15, making hearing protectors essential during the working day. Keywords: ergonomics; acoustic comfort; agricultural mechanization; sound pressure level.