ABSTRACT Active optical sensors have been widely used for the spatial and temporal monitoring of peanut culture because they are accurate, non-destructive methods for rapidly obtaining data. The objective of this study was to determine the optimal stage of crop growth for collecting sensor readings based on correlations between quality indicators. In addition, we compared vegetation indices (Normalized Difference Vegetation Index [NDVI], Normalized Difference Red-Edge Index, [NDRE], and Inverse Ratio Vegetation Index, [IRVI]) by monitoring temporal variability in the peanut crop in order to determine which of them obtained the best reading quality throughout the process. The experiment was performed on the 2016/17 crop in the agricultural area of the municipality of Dumont in the state of São Paulo, Brazil. The experimental design was based on the basic assumptions of statistical quality control and contained 63 sample points in a 30 × 30 m grid. The parameters were evaluated at 30, 45, 60, 75, and 119 days after sowing (DAS) using proximal sensing with GreenSeeker and OptRX sensors. We found that 45 and 60 DAS were the optimal times for monitoring peanut crop variability. For spatiotemporal monitoring of the culture with control charts, NDRE showed the best readings throughout the process when compared to NDVI and IRVI.

RESUMO A qualidade da colheita mecanizada de soja é essencial para garantir que todo investimento realizado nas demais etapas do processo produtivo da cultura seja convertido em lucro para o produtor. Considerando que a qualidade dessa operação está relacionada com menores perdas e que a plataforma de corte utilizada e a velocidade de colheita podem influenciar nisso, objetivou-se avaliar a qualidade da colheita mecanizada da soja, no Cone Sul de Rondônia, em colhedoras com diferentes plataformas de corte e velocidades de deslocamento por meio do controle estatístico de processo. O experimento foi realizado na safra 2015/16 em uma área agrícola no município de Cabixi, RO, localizada nas coordenadas geográficas 13º 27’ S e 60º 41’ W. O delineamento experimental foi estabelecido de acordo com as premissas do controle estatístico de processo, sendo amostrados 80 pontos distanciados a cada 100 m. Avaliaram-se as perdas na plataforma, nos mecanismos internos e totais e a altura de corte. A melhor qualidade do processo de colheita foi obtida na plataforma de esteira, com pouca influência da velocidade de deslocamento. O nível de perdas na colheita foi acima do limite aceitável, indicando necessidade de melhorias do processo.

ABSTRACT Assuring the quality of mechanical soybean harvesting is crucial to ensure that investment in other stages of crop production is converted into profit for the farmer. Because the quality of this operation is related to decreased losses, which may be affected by the combine header used and by the harvesting speed, the aim of the present study was to assess the quality of mechanical soybean harvesting in the Southern Cone of Rondônia, using harvesters with different combine headers and travel speeds via statistical process control. The experiment was performed during the 2015/16 harvest in an agricultural area of the municipality of Cabixi, RO, located at the geographical coordinates 13° 27’ S and 60° 41’ W. The study was designed based on the premises of the statistical process control, sampling 80 points 100 m apart. Combine header, internal, total, and cutting height losses were assessed. Quality of the harvesting process was assessed using the draper header. Travel speed had little effect on process quality. Crop losses were higher than the acceptable limit, indicating the need to improve the process.