Avaliação de lavouras de café por índices de imagens multiespectrais obtidas por Aeronave Remotamente Pilotada

Abstract

Brazil stands out in worldwide coffee production. In Minas Gerais, São Paulo and Espírito Santo states the production represents more than three quarters of the country's total. Precision coffee farming is an effective approach to keep the country with high production numbers and increase producers competitiveness. Remote sensing and Unmanned Aerial Vehicles (UAVs) are able to improve the way coffee is managed, favoring a reduction in the use of agricultural inputs and in environmental liabilities, increasing profitability and product quality. In this context, the objective of this work was to evaluate the effectiveness of two different types of management implemented in the coffee tree, through high resolution multispectral images, obtained by UAV and vegetation indexes (VI). The study was carried out at Fazenda Samambaia in Santo Antônio do Amparo, MG, where two different agricultural practices were adopted as objects of study in this work and that were confronted with separate areas as control, which represented the conventional for coffee farming: deep liming (between 0.60 m and 0.80 m) vs. limestone application in the projection of the coffee canopy; planting seedlings produced in tubes vs. seedlings produced in multihole polyethylene bags. For this evaluation, bimonthly UAV flights were performed with subsequent processing of the images obtained. Measurements of plants height, crown diameter and chlorophyll content were conducted in field. The VI results were compared to the other field measurements, via temporal graphics and through correlation analyzes. Linear models were also developed to predict the biophysical parameter best correlated with VIs. In addition, for the study of planting seedlings produced in tubes vs. multihole bags, the percentage of failures resulting from seedlings not “taking on” was assessed. As a result of the deep liming study, a significant difference was found only for the chlorophyll parameter, that obtained a higher mean value in the section where the management was carried out (total chlorophyll of 86.04 μg / cm² against 83.87 μg / cm²). In the study involving different containers (tubes vs. multihole bags), there was no significant difference between the parameters evaluated. However, in the failures assessment, it was found that the area with seedlings produced in multihole bags obtained a significantly lower result (6.4% failure versus 11.7% for the tubes). Most of the indices had a high correlation with field parameters except for chlorophyll, with emphasis on the GNDVI, NDVI, GCI and vectorized canopy area (cm²). For both experiments, the last two obtained values close to 0.90 with the Leaf Area Index (IAF). It was possible to develop linear models for the prediction of biophysical parameters, such as the LAI, as a function of the GCI for both assessed managements, with R² values above 0.75. The UAV proved to be an effective tool to monitor the coffee crop and to evaluate management practices.