Ajuste de modelo para análise de porcentagem do índice de vegetação da diferença normalizada adaptado (ndvi*) obtidas por veículo aéreo não tripulado (vant) em pastagem de aveia Ucraniana (avena strigosa schreb.)

Abstract

The productive capacity of the national primary sector places the country as a great potential for the supply of these world demands. In this context, Brazil proposes, with the creation of the ABC Plan (Low Carbon Agriculture), the Crop-Livestock-Forest Integration (ILPF), a strategy that aims to integrate forest components, annual crop and pasture species, and livestock. The objective of this work, with the use of UAV, was to propose a model adjustment to evaluate the possibility of obtaining the vegetation cover index, through the Adapted Normalized Difference Vegetation Index (NDVI*), to serve as a subsidy for decision making about the management of an area cultivated with pasture, conducted under rotational grazing management. The work was carried out on a property located in Rosário do Sul - RS. Field surveys were carried out with a Mavic 2 Zoom (DJI) model drone, image processing in Agisoft Metashape Pro software, and orthophoto processing in QGIS software. R Studio was used for clustering using K-means, considering K=5. The analyzes carried out in Survey 1 show that P1 has a higher percentage of exposed soil compared to the other paddocks, and the highest percentage of vegetation available for the herd is given by P3. In Survey 2, it can be seen that P2 had the highest percentage of exposed soil, on the other hand, P3 had the highest percentage of vegetation cover. According to Survey 3, it can be seen that P3 had the highest percentage of exposed soil in relation to the other paddocks. P4 had the lowest exposed soil value, indicating the highest vegetation cover rate. According to Survey 4, P4 showed a higher percentage of exposed soil. However, P3 showed a higher percentage of vegetation cover in relation to the other paddocks. This method proved capable of estimating the percentage of vegetation cover in the soil using the NDVI*, with this, this methodology can be used to support the decision making about rotational grazing.