Obtenção de imagens digitais e espectrometria da cultura do milho e sua relação com produtividade em diferentes doses de nitrogênio

• Ano de defesa: 2014
• Autor(a) principal: Gasparotto, Aline de Carvalho
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual de Maringá; Brasil; Departamento de Agronomia; Programa de Pós-Graduação em Agronomia; UEM; Maringá, PR; Centro de Ciências Agrárias
• Palavras-chave em Português: Solo; Milho; Indices de vegetação; Espectroradiômetro; Câmera fotografica digital; Nitrogênio; Brasil.; Corn; Indices of vegetation; Spectroradiometer; Digital photographic camera; Nitrogen; Brazil.; Ciências Agrárias; Agronomia
• Link de acesso: http://repositorio.uem.br:8080/jspui/handle/1/1272

The monitoring of agricultural production in Brazil is a complex activity, mainly due to the large expanse of planted areas and its difficult access. Therefore, the development of techniques that assist in the identification of nutritional stress, and yield components are important. The use of reflectance data obtained by the spectroradiometer, enables to establish patterns of spectral response of crops, enabling the introduction of methodologies for surveying and monitoring of the same. The objectives of this work were to propose vectors vegetation indices extracted from digital images captured by a digital camera using different filters ( IR 720nm , IR 850nm and UV-IR cut ) comparing with the rates obtained with the use of a spectroradiometer; evaluate the use of the indices alone in adjusting for linear discrimination between levels of nitrogen in corn models, as well as its relationship with yield components, and to establish the spectral behavior of maize to different nitrogen levels applied during its development. Methodologically the experiment was prepared in the greenhouse and it was a corn plant per pot. The design was a completely randomized design with 5 treatments with nitrogen rates of 0, 50, 100, 200 and 300 kg ha-1, with 10 repetitions . Evaluations were performed at 15, 30 , 45 , 60 and 80 days after emergence (DAE), with the capture of images that covered the range of the visible and near infrared (350-1000 nm) using a series of filters in a Fujifilm IS Pro digital camera, plus the readings with the spectroradiometer. The data collected were processed using the Spring and Excel software to calculate the vegetation indices GNDVI, NDVI, GNIR and RNIR . At 80DAE evaluated to dried mass, fresh mass, leaf area and leaf nitrogen. As a result, it was observed that the spectral reflectance is changed from corn with different doses at 60 and 80 DAE, mainly in the visible region of the spectrum, whereas plants showed a higher reflectance with low doses of N. The GNDVI and GNIR indices showed higher sensitivity for assessment of nitrogen deficiency in corn. The use of camera with different filters is presented as a promising tool for the discrimination of nitrogen, requiring further studies to decrease the coefficient of variation of the samples. The vegetation indices studied using the 720nm filter was superior to 850nm in correlations with indices spectroradiometer. All vegetation indices and yield components which were significant in the regression showed a quadratic behavior, with the exception of leaf nitrogen that obtained linear behavior . The best seasons for discrimination of nitrogen by vegetation indices were evaluated at 60 and 80DAE, and production that over correlated with the vegetation index component was the leaf area.