Banca de DEFESA: VITÓRIA CÔRTES DA SILVA SOUZA DE OLIVEIRA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : VITÓRIA CÔRTES DA SILVA SOUZA DE OLIVEIRA
DATE: 28/05/2024
TIME: 10:00
LOCAL: Google meet
TITLE:

ESTIMATION OF TIFTON85 FORAGE ATTRIBUTES USING REMOTELY PILOTED AIRCRAFT

KEY WORDS:

Remote sensing, digital images, forage crops, vegetation index.

PAGES: 68
BIG AREA: Ciências Agrárias
AREA: Engenharia Agrícola
SUBÁREA: Máquinas e Implementos Agrícolas
SUMMARY:

With technological advances and the need to increase the efficiency of the production process, precision agriculture emerges as an alternative for better production planning and management, as it provides infinite benefits with good productivity, sustainability, and economic development. In this sense, remote sensing techniques have been widely used instantly and at low cost, playing an important role in diagnoses such as estimating productivity and crop attributes. The objective of this research was to evaluate the spectral responses of tifton85 at different growth stages and generate models capable of estimating forage attributes using remote sensing techniques. The experiment took place over an interval of 48 days, with routine collections totalling 5 evaluation periods (EA). In each EA, the following attributes were collected at 39 georeferenced points: leaf area index (LAI) using the AccuPAR L-80 ceptometer, plant height with a millimeter ruler and chlorophyll content with the Falker chlorofiLOG CFL1030 chlorophyllometer. Concurrently, images were obtained with a DJI Phantom 4 Advanced remotely piloted aircraft with a multispectral sensor on board and vegetation indices (VIs) were calculated for each point in each EA. At the end of this experiment period, biomass was collected at each point. Pearson's correlation was calculated between crop attribute data and IVs and the highest correlation index was performed using a simple linear regression to generate attribute estimation models. The models generated only with IVs obtained: chlorophyll (R² = 0.7202), height (R² = 0.5744) and IAF (R² = 0.7539). Principal component analysis was used to investigate the structure of the data and identify underlying patterns, with the aim of reducing the set of information and obtaining the most relevant indices. The set of main components with accumulated explanatory power greater than 70% of the data variance was considered relevant for reducing variable sizing and was called the crop growth index (ICC). A linear regression was performed using the ICC to generate attribute estimation models based on the IVs in a multivariate manner. The models generated with ICC obtained: chlorophyll R² = 0.7009, height R² = 0.5336 and IAF R² = 0.6582. In the case of biomass, the data was grouped into 8 points. The IVs obtained at each time were correlated with the final biomass values. Degree 2 polynomial regression was performed with the 3EA GNDVI index (r = -0.82298) predicting the final biomass of the crop, obtaining R² = 0.8078.

COMMITTEE MEMBERS:
Presidente - 2161955 - ANDERSON GOMIDE COSTA
Interno - 1050615 - MURILO MACHADO DE BARROS
Externo à Instituição - RAFAEL DE OLIVEIRA FARIA - UFLA