Modelagem da produção de biomassa da aveia preta (Avena strigosa Schreb.) através de imagens obtidas por aeronave remotamente pilotada

Abstract

Agricultural monitoring becomes a determining factor when the goal is to increase the productivity and profitability of a crop. Currently one of the most used techniques for monitoring agricultural areas is Remote Sensing. From the use of images, it is possible to obtain estimates of productivity, control diseases and pests, and evaluate the water requirement of plants, thus providing important subsidies to decision making. It is in context that the use of Remotely Piloted Aircraft System (RPAS) emerges as an important option, given the remarkable capacity of these platforms to provide data with highest spatial and temporal resolution, allowing a more effective Remote Sensing in small scale applications, such as agriculture. In this sense, the objective of this work was to evaluate the potential of using images acquired through a RPAS platform to obtain three dimensional models representing the canopy of the black oat crop. From these models, plant height, fresh biomass and dry biomass were estimated, evidencing the spatial variability of these attributes. For data collection, a test site was evaluated at the Universidade Tecnológica Federal do Paraná (UTFPR), Dois Vizinhos campus. The collections occurred between July and August of 2017, where a RPAS platform with RGB sensor was used on board. Flight planning was carried out in a smartphone application, where a lateral and longitudinal overlap between images of 90 % was defined, flight height at 25 meters, and the delimitation of the area to be covered. From this information, along with the global navigation system (GNSS) receiver in the equipment, the flights were realized in an autonomously . After the image processing, digital surface models (MDS) with spatial resolution of 1.8 cm.pixel-1 were obtained, duly associated to control points in the field, which ensured the positional accuracy of the data. In parallel, data were collected in the field, where the height of plants, freshmatter and dry matter production were measured for each of the plots evaluated. Alongwith the database from the RPAS, a series of regression analyzes were performed, in which mathematical models were obtained to estimate the variables studied. As a result, R² was obtained between 0.77 and 0.91 for plant height derived from RPAS and height of plants measured in the field. On the other hand, the analysis of biomass production resulted in a R²between 0.64 and 0.94 for the different missions, with a relative error of less than 17%. The models obtained here demonstrated the high potential of the use of RPAS as a way of estimating plant height and biomass production for black oats, allowing the tool to be used in the monitoring of these areas, providing an important contribution to decision making.