ABSTRACT Macaw palm (Acrocomia aculeata) is a potential raw material for biodiesel production owing to its high oil content. In this study, a model is created that can estimate the oil percentage in Macaw fruit using spectral indices obtained via computer vision. Ninety Macaw fruits are used to perform an experiment. They are categorized into three stages of maturation and counted in terms of weeks after flowering. Images of the fruits are captured using a multispectral camera in the visible spectral region (RGB) and infra-red regions (NIR), which allows spectral indices to be obtained. A model is established using the principal component regression method and used to estimate the oil content based on the colorimetric characteristics of Macaw palm fruits. The selected model is composed of the first two principal components, which indicate an explanatory potency of 91.57% of the variance. Applying this information to the validation data, the model exhibits a determination coefficient of 0.91, mean square error of 10.61%, and standard error of 5.46% in terms of the oil content, which indicates that computer vision is a promising alternative method for estimating oil content.

ABSTRACT Macaw palm (Acrocomia aculeata) is a promising crop for biofuel production due to the high concentration of its fruit oil, but the harvest date is an issue to be better understood so it could be cultivated on an industrial scale. The aim of this study was to use the colorimetric properties of the macaw palm fruits to develop a neural network classifier to determine the ideal moment for harvesting, based on the oil content of the fruit mesocarp. During nine weeks of maturation were sampled 900 fruits of macaw palm fruits and the colorimetric properties of the RGB, HSI and CIELab color models were used to classify the fruits into immature and mature fruits. Kappa index and the overall accuracy values were used to access the classifier performance. The classifiers based on RGB parameters and on hue were considered equivalents having a Kappa index of 0.901 and 0.942, respectively, indicating the 59th week of maturation as the ideal time to harvest with the highest oil content.

Os objetivos deste trabalho foram avaliar o comportamento agronômico, quantificar a variabilidade e estimar as distâncias genéticas de 66 acessos de videira destinadas à elaboração de vinhos, presente no Banco de Germoplasma de Videira da EMBRAPA Semiárido, em Juazeiro-BA, Brasil, por meio da caracterização de descritores fenotípicos de variação contínua e discreta. As técnicas multivariadas utilizadas, componentes principais, método de otimização de Tocher e projeção gráfica das distâncias foram eficientes no agrupamento dos genótipos mais similares, de acordo com as suas características fenotípicas. Não houve concordância na formação dos grupos pelo método de otimização de Tocher, quando foram avaliadas características morfoagronômicas de variação contínua e discreta. A utilização de variáveis discretas permitiu a separação de Vitis vinifera e híbridos em grupos distintos. Correlações significativas positivas foram observadas entre peso, comprimento e largura de cachos, bem como correlação negativa entre acidez total titulável e relação SST/ATT. 84,12% da variação total presente nos dados originais foram explicados pelos primeiros quatro componentes principais. Os resultados obtidos demonstram que existe pequena variabilidade entre os acessos de uvas para vinho no Banco de Germoplasma da Embrapa Semiárido.

The objectives of this study were to evaluate the performance of cultivars, to quantify the variability and to estimate the genetic distances of 66 wine grape accessions in the Grape Germplasm Bank of the EMBRAPA Semi-Arid, in Juazeiro, BA, Brazil, through the characterization of discrete and continuous phenotypic variables. Multivariate statistics, such as, principal components, Tocher's optimization procedure, and the graphic of the distance, were efficient in grouping more similar genotypes, according to their phenotypic characteristics. There was no agreement in the formation of groups between continuous and discrete morpho-agronomic traits, when Tocher's optimization procedure was used. Discrete variables allowed the separation of Vitis vinifera and hybrids in different groups. Significant positive correlations were observed between weight, length and width of bunches, and a negative correlation between titratable acidity and TSS/TTA. The major part (84.12%) of the total variation present in the original data was explained by the four principal components. The results revealed little variability between wine grape accessions in the Grape Germplasm Bank of Embrapa Semi-Arid.