Correlation between vegetation indices obtained by rpa and grain yield

Resumo

The objective of this study was to evaluate the correlations between vegetation indices (VI), obtained by RGB camera on flights of Remotely piloted aircraft System (RPA), with yield maps of agricultural crops. Monitoring was carried out during 4 harvest seasons: soybean 2018/19, maize (2019), soybean 2019/20 and wheat (2020), in two areas of a rural property located in Toledo, Paraná. During the harvests, periodic flights were performed using DJI-branded RPA Phantom 3 Advanced. For the generation of orthomosaic, Agisoft PhotoScan software (Free trial) was used. After the RGB bands normalization, the vegetation indices MPRI, VARI, GLI and ExG were calculated for 3 flight dates in each harvest in the study areas. At the end of the crop cycle, samples were collected to create the yield maps. With the yield data, descriptive statistics analyzes were performed and, later, the correlation between the VIs and the yields of each harvest was performed using Spearman's correlation coefficient (rs). According to the research, it would be suggested that the farmer carry out surveys with RPA with RGB camera in soybean crop, mainly in R7 stage, in maize at VT (bolting) stage and in wheat at tillering stage, since these phenological stages showed higher correlations and between the VIs and the yield of each crop. The pairs of VIs MPRI and VARI, GLI and ExG were similar as vegetative indicators, so only two of them would already have the capacity to represent the variations existing in the areas of the study between the dates.