ABSTRACT Brazil is currently the 4th world’s largest banana producer, producing around 7 million tons. In this scenario, several studies have been developed with a large amount of data, such as climatic, morphological, and nutritional data, in an attempt to improve these numbers even further. This study aims to classify banana ripening stages by artificial neural networks (ANN) as a function of plant physical, physicochemical, and biochemical parameters. The used ANN consisted of a three-layer feedforward backpropagation network, with eight neurons in the input layer (physical, physicochemical, and biochemical parameters), ten neurons in the intermediate layer, and two neurons in the output layer (classification of banana ripening stages). The results showed three configurations. ANN presented an excellent result for the training phase, with 100% accuracy in the sample classification for the three configurations. The validation and testing phases, that is, the classification of samples that were not part of the training, showed 91.6% and 94.4% accuracy in the first and second configurations, respectively, and 89.5% accuracy in the third configuration.

RESUMO No processo de recuperação de um solo dois pontos são fundamentais: definir as melhores interferências quanto à aceleração do seu processo de reabilitação e os indicadores mais adequados para diagnosticar sua qualidade. Portanto, este trabalho teve como objetivo investigar as nanopartículas e a morfologia da nanoestrutura de um Latossolo Vermelho em recuperação há oito anos, como indicadoras da sua qualidade. Comparou-se o solo em processo de recuperação com seu estado natural e degradado. Foram quantificadas as nanopartículas (ø < 100 nm) e a argila fina (ø < 200 nm), assim como a argila total e areia, e estudada a morfologia de nanoestruturas por meio de imagens obtidas por microscópio eletrônico de transmissão. Para auxiliar na interpretação da qualidade do solo foram analisados os atributos físicos (porosidade e densidade do solo) e químicos (C, nitrogênio total e hidrogênio). As nanopartículas e a morfologia de nanoestruturas do solo são bons indicadores da sua qualidade; os atributos físicos e químicos estudados não são sensíveis para detectar alterações entre as condições de solo degradado e em recuperação após oito anos; na classe de partículas com diâmetro menor que 200 nm é mais eficaz a visualização de nanoestruturas.

ABSTRACT In the process of rehabilitation of a soil, two points are fundamental: to define optimal interferences to accelerate the rehabilitation process and the most appropriate indicators to diagnose its quality. Therefore, this study aimed to investigate the nanoparticle and nanostructure morphology of a Red Latosol in rehabilitation for eight years. The soil under rehabilitation process was compared with its natural state and degraded. In the topsoil, nanoparticles (ø < 100 nm) and fine clay (ø < 200 nm) were quantified and the nanostructures morphology was studied using images obtained by transmission electron microscopy. Soil porosity, bulk density and carbon, nitrogen and hydrogen contents were analyzed. It was found that the nanoparticles and nanostructure morphology were good soil quality indicators; the physical and chemical attributes were not sensitive to detect alterations between the conditions of degraded soil and soil rehabilitated for eight years; in the class of particles with diameter smaller than 200 nm, for the studied Red Latosol, the visualization of nanostructures is more effective.