Sensoriamento remoto para mapear o potencial de produtividade do cafeeiro

Abstract

The use of satellite images is becoming a powerful tool for the management and trade of agriculture products, both for the producer and government funding agencies. Even though the potential of use is high, there are no researches that estimate, with high accuracy, the productivity of coffee crops based on the vegetation index. The objective of this study was to correlate the potential of coffee crops productivity with normalized difference vegetation index (NDVI, calculated from Sentinel-2 satellite images. The trial was carried out in three crops of same age, cultivar and plants spacing located in Minas Gerais at the city of Santo Antônio do Amparo. The trees were cut at nearly 30 cm above the ground in 2015 and they were harvested in 2017 and 2018. The sample was defined at each 0,5 ha, and inside this area four trees were harvested. The harvest was performed manually, and the harvested volume was measured by a liter graduated bowl (liters.plant-1). The vegetation indexes were calculated using Sentinel-2 Satellite multispectral images obtained in different months from 2016 trough 2018. The images were imported into software R Studio, where the whole process to calculate the NDVI and its productivity correlation were made. Four different ways to correlate NDVI and productivity in two years were tested, two of them considering the results of each year, and the other two considering the results of the two years together, both of them grouping the results point by point and by average. As a result, it can be said that the best timing for the image acquisition corresponds to the months of August and September, and the methodologies that used average results for each of the areas, as much for years apart, as for two years together, were the ones that presented the highest values of R². The point-to-point methodology for each year of evaluation was not satisfactory, but when evaluated in 2 years it presented excellent R2 values. Therefore, the use of Sentinel-2 Satellite images to get the NDVI is a powerful tool to estimate the productivity potential of coffee crops.