Índices de vegetação em lavouras cafeeiras por sistema de aeronave remotamente pilotada

Abstract

Brazil is the main producer, exporter and second-largest consumer of coffee in the world and Minas Gerais State with great contribution to the national economic scenario. The application of technologies in the agricultural sector, therefore, optimization in the study, research and intelligent decision making ensuring greater profitability of crops. Thus, the use of electronic devices with embedded sensors to capture aerial photographs of crops, with emphasis on the Remotely-Piloted Aircraft System (RPAS), demonstrates such applicability, through the use of Remote Sensing techniques. In this context, it was initially proposed to survey information of conceptual character and of works developed in the sectors of study to build a theoretical reference for the realization of the proposed objectives. In sequence, the first study proposed to characterize three recently planted coffee cultivars, analyzing their temporal behavior, evidenced by the action of the dry and rainy periods in the development of the plants in the first year of formation in the field, as well as describe the behavior of the spectral profile of the three cultivars for the two periods of study, statistically differentiate the cultivars using the studied variables and estimate linear regression equation between the radiometric data of the vegetation indexes (VIs) and the total chlorophyll data (Chl t) and leaf area index (LAI). The experiment was conducted at Samambaia Farm, located in the municipality of Santo Antônio do Amparo, Minas Gerais, with recently planted coffee (Coffea arabica L.) cultivars Catucaí (2SL), Catuaí (IAC 62) and Bourbon (IAC J10) and aged 5 months at the beginning of the work. The images were obtained through RGB and multispectral sensors in Remotely-Piloted Aircraft System (RPAS), collected every two months, from May 2019 to March 2020, as well as height data, crown diameter, chlorophylls, and vegetation indexes (VIs). According to the results obtained, it was possible to characterize and verify differences between the study periods, except for the chlorophyll variable. Significant statistical differences were detected that distinguish for the rainy period the Bourbon cultivar from the Catucaí and Catuaí cultivars. The spectral characterization showed proximity and overlap between the spectra of the cultivars for both periods of study, and it was not possible to individualize the coffee cultivars. It was possible to estimate the model and linear regression equation only for the LAI for the three cultivars studied highlighting the applicability of data obtained by RPAS in precision coffee farming studies.