ABSTRACT Analyses of fertilizer are essential to ensuring that fertilizer sold to final users presents chemical and physical qualities in the range determined by law. As regards the sampling of fertilizers, the official method currently used in Brazil for sampling preparation is to reduce the size of the sample (from ~ 3 kg to ~ 0.25 kg) by quartering, followed by grinding (so as to pass through a 0.85 sieve) and nutrient quantification. Herein, we propose an alternative method of sampling preparation by grinding the total sample (~ 3 kg) before quartering to improve accuracy and reduce segregation during quartering. Six formulations of fertilizers (basal samples) were weighed (0.01 kg precision) and sampled according to the two methods (official and alternative), followed by the quantification of nutrient concentration in duplicate. Results showed that both methods presented similar nutrient concentrations for most formulations compared to the basal samples. However, the alternative method presented higher precision (less variation between replicates) and accuracy (versus the basal samples) than the official method. Consequently, the alternative method can be used for sampling preparation fertilizers with high accuracy and precision in determining nutrient concentration. law from 025 0 25 0.2 so 085 85 0.8 sieve Herein ( samples 0.01 001 01 (0.0 alternative, , alternative) duplicate However less replicates versus Consequently 02 2 0. 08 8 0.0 00 (0. (0

RESUMO A análise de imagens pode fornecer informações extraídas das folhas das culturas, que podem contribuir na identificação precoce de deficiência de nutrientes. O objetivo do trabalho foi reconhecer padrões nutricionais de nitrogênio (N) em plantas de milho, nos estádios V4 e V7, utilizando análise de imagens digitais baseados em índices espectrais. O experimento foi realizado em casa de vegetação sob cultivo hidropônico. Os tratamentos foram dispostos em um delineamento inteiramente casualizado, em esquema fatorial 4×2, com quatro repetições. Os fatores foram constituídos pelas doses de N (0; 3,0; 6,0 e 15 mmol L-1) combinados no V4 e V7. Em cada estádio foram digitalizadas imagens da lâmina foliar, que foram processadas pela análise de imagens e analisadas quimicamente. Para reconhecimento das imagens foi utilizado um vetor de características baseados nos índices espectrais: excesso de verde, vermelho normalizado, verde normalizado e razão verde-vermelho, e na combinação entre eles. Na imagem original foram extraídos blocos de 9×9; 20×20 e 40×40 pixels. Para validação dos resultados foi determinado o teor de N na lâmina foliar, a massa seca das plantas e o nível crítico externo de N na solução nutritiva, com base em 90% produção de massa seca. A máxima porcentagem de acerto global dos padrões de N foi 80 e 93% em V4 e V7, respectivamente. A utilização de índices espectrais combinados proporcionou melhor desempenho na classificação, e o bloco de imagem 20×20 pixels mostrou-se mais adequado para a diferenciação entre as doses de N.

ABSTRACT Image analysis can provide information extracted from the leaves of crops, and contribute to early identification of nutrient deficiency. The objective of this study was to recognize nutritional nitrogen (N) patterns in maize plants, at the V4 and V7 stages, using digital image analysis based on spectral indexes. The experiment was carried out in a greenhouse under hydroponic cultivation. Treatments consisted of a completely randomized design, in a 4 × 2 factorial arrangement, with four replications. The factors were constituted by the doses of N (0; 3.0; 6.0 e 15 mmol L-1) combined at V4 and V7. In each stage, digital images were taken of leaf blades, with subsequent chemical composition and image analysis. For image recognition and classification, a vector of characteristics based on the spectral indexes was used as follows: excess of green, normalized red, normalized green and red-green ratio, and the combination among them. Additionally, extracted blocks of 9 × 9, 20 × 20 and 40 × 40 pixels on original images were used. The N content in the leaf blade, the dry mass of the plants and the external critical level of N in the nutrient solution were determined for result validation, based on 90% dry matter production. Maximum the global accuracy rate for N patterns was 80 and 93% at V4 and V7, respectively. The use of combined spectral indexes provided better classification performance, and the 9 × 9 pixel image block appeared more adequate for differentiation among the doses of N.