ESTIMATIVA DA PRODUTIVIDADE DA SOJA POR MEIO DO SENSORIAMENTO REMOTO PROXIMAL POR MEIO DOS ÍNDICES DE VEGETAÇÃO NDVI E NDRE
ESTIMATIVA DA PRODUTIVIDADE DA SOJA POR MEIO DO SENSORIAMENTO REMOTO PROXIMAL POR MEIO DOS ÍNDICES DE VEGETAÇÃO NDVI E NDRE
-
DOI: https://doi.org/10.22533/at.ed.738122516122
-
Palavras-chave: agricultura de precisão, Glycine max (L) Merril, sensoriamento remoto, sensor óptico ativo, agricultura sustentável.
-
Keywords: sustainable agriculture, precision agriculture, Glycine max (L.) Merril, remote sensing, active optical sensor.
-
Abstract: Remote sensing (RS) is a valuable geotechnological tool that facilitates the implementation of sustainable agricultural practices. As a non-destructive method, it allows for the monitoring of temporal and spatial variability in crops, enabling more informed and timely decision-making. Based on this premise, our study explored how sowing density impacts soybean productivity through spectral monitoring of the crop, utilizing proximal remote sensing (PRS). The objective was to estimate soybean productivity as a function of sowing density and to identify the optimal times for evaluation using vegetation indices. This research was conducted in an experimental area at the Federal Technological University of Paraná (UTFPR), Santa Helena Campus, in the state of Paraná. The experimental design was completely randomized in strips, consisting of two treatments, two sowing densities (21.2 and 22.5 seeds m⁻¹), and four replicates. The variables analyzed were vegetation indices (NDVI - Normalized Difference Vegetation Index and NDRE - Normalized Difference Red Edge Index) and productivity. The periods evaluated for vegetation indices were 36, 69, 76, 85, and 92 days after sowing (DAS). Productivity was evaluated only at the end of the crop cycle, when soybeans reached physiological maturity. Statistical analyses included boxplots, Pearson's correlation coefficient, and polynomial model regression to estimate yield, using the coefficient of determination (R²) as a measure of accuracy. The analysis concluded that the NDRE provided better results for yield estimation compared to the NDVI, which exhibited issues with saturation. For a sowing density of 21.2 seeds per meter, the optimal times for estimating productivity were at 76 and 92 DAS. In contrast, for a sowing density of 22.5 seeds per meter, the best times for estimation were at 85 and 92 DAS.
- Sarah Wachtel
- Joseane Aparecida De Paula Cardoso
- Henrique Carlos Mognol
- Matheus Gabriel Anschau
- Vitor Huggo Arenhartt Radetzki
- Gabriele Vitoria Silva Serena
- Bárbara Carolline Matwijou
- Gesivaldo Ribeiro Silva
- Armando Lopes De Brito Filho
- Lincon Oliveira Stefanello
- Maura Gabriela Da Silva Brochado
- Franciele Morlin Carneiro