Predição da produtividade de milho (zea mays l) por meio da interpretação temporal de índices de vegetação

• Autor(a) principal: Meirelles, Luara Franciane de
• Data de Publicação: 2022
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/15993
• topic: NDRE; NDVI; Produtividade; Sensoriamento Remoto; Remote Sensing; Productivity; CIENCIAS AGRARIAS::AGRONOMIA

The estimation of grain production is necessary for agribusiness. In Brazil, the corn culture (Zea mays L.) has become one of the leading agricultural commodities, being the basis for the country's agribusiness. Brazil's biggest region corn producer is the Center-South, which represents 83% of national production. The use of remote sensing can generate an extensive database, helping the producer manage and monitor the crop and predict productivity. An example of remote sensing combined with monitoring and predicting productivity is through techniques that relate productivity with vegetation indices obtained from satellite images. This work aimed to evaluate methods for estimating the yield of second-crop corn on a property in the Southwest region of São Paulo using vegetation indexes. For this purpose, we used the Normalized Difference Vegetation Index (NDVI), Normalized Difference Vegetation Index (NDRE), and data production obtained directly from a productivity sensor installed in the harvester John Deere S660 model. The study area is located on a property in the municipality of Capão Bonito-SP, in an area of 19 ha of second-crop corn. The values of vegetation indices were obtained through Sentinel-2 satellite images (pixel 10×10m) and using the QGIS software. In total, 15 scenes between planting and harvesting were acquired. We used the average productivity values obtained within each pixel, thus correlating the values of vegetation indices with the productivity value within the pixel. The linear regression calculation of productivity generated 30 graphs, 15 for each index with the machine's productivity. In this study, it was observed that the most accurate linear regression was using NDVI on June 13, 2020, which corresponds to 100 days after sowing. Therefore, it can be seen that through this technique, it is possible to monitor corn productivity before harvesting, offering an easier way to obtain data and manage the crop.