Estimate of Vigor Classes of Brachiaria Ruziziensis using Sensors Boarded on UAV Platform

Abstract

Traditional procedures for biomass estimation usually use destructive methods with great demands on time, resources, and labor. The development of models for automated estimation of pasture biomass, particularly from images captured by Unmanned Aerial Vehicle (UAV), in addition to high spatiotemporal resolution combined with flexibility in image acquisition, provides agility, the economy of resources, and labor. The objective of this work was to establish a technical feasibility study for the use of multispectral sensors onboard an Unmanned Aerial Vehicle (UAV) to estimate the vigor classes of Brachiaria ruziziensis pastures. For this purpose, imaging cameras in the visible (RGB), near-infrared and red edge ranges were used for continuous monitoring of 20 pasture paddocks with an area of 1,350 m2 each, totaling 27,000 m2 of the experimental area. The indices performed well and were sensitive in class discrimination at intervals that range from soil exposure and stresses caused by pest and disease infestation (low vigor) to conditions in which the vegetation is in good development, in class intervals with high levels of vegetation and, consequently, pointing to high values of biomass.