Uso de veículo aéreo não tripulado para análise espectral e espacial da fitossanidade em plantios de cana-de-açúcar na Amazônia oriental.

Abstract

Unmanned Aerial Vehicles (UAVs) have become an important innovation tool in different areas of society and science, and can be used in the most diverse situations, improving and optimizing results. The broad field of Precision Agriculture (AP) appears as one of the sectors that best integrates this technology, being used in the most different agricultural species, such as, for example, in the cultivation of sugar cane, which will be studied in this work. According to what was reported, the objective of this research was to use UAV platforms for spectral and spatial analysis of sugarcane plantations in areas located in the Brazilian Eastern Amazon, a region lacking in research using the proposed tool and in relation to the culture studied. The research was subdivided into four chapters. In the first chapter, a contextualization of sugarcane cultivation and the Brazilian Eastern Amazon was carried out, as well as a statistical analysis in relation to planted areas, productivity and production of the crop, over ten subsequent harvests, seeking to understand the difficulties and potential of the region. The second chapter consisted of a bibliographical review on the use of UAVs in sugarcane fields in Brazil, seeking to demonstrate where, how much and how this innovation has been used in national sugarcane fields. In the third chapter, the processing of images obtained by the UAV began, making it possible to carry out statistical-spectral analyzes of the vegetation indices NDVI (Normalized Difference Vegetation Index), SAVI (Soil-Adjusted Vegetation Index), NDRE (Normalized Difference Red Edge Index) and GNDVI (Green Normalized Difference Vegetation Index), in the sugarcane areas studied, allowing the definition of the index that obtained the best fit to the data, through correlation and regression analyses. Finally, the fourth chapter was the last to be discussed in this research, making it possible to evaluate different spatial interpolators to define phytosanitary management zones in areas with sugarcane plantations, in order to guide the management of potential infestations of the legume black mucuna, considered a weed for sugarcane. The interpolators used were ordinary kriging (KO) and suppot vector machine (SVM), coming, respectively, from geostatistics and machine learning. At the end of this chapter, two annexes were inserted, one for each interpolator used, referring to the results of carrying out a literature review to understand the current state of research with both interpolators in sugarcane plantations in Brazil.