Aplicação de sensor multiespectral a bordo de VANT na determinação de graus de severidade de Erwinia psidii em Eucalyptus urograndis

• Ano de defesa: 2017
• Autor(a) principal: Pedrali, Letícia Daiane
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• dARK ID: ark:/26339/001300000504b
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Recursos Florestais e Engenharia Florestal; UFSM; Programa de Pós-Graduação em Engenharia Florestal; Centro de Ciências Rurais
• Palavras-chave em Português: Sensoriamento remoto; Eucalyptus urograndis; Índices de vegetação; Erwinia psidii; Remote sensoring; Vegetation rates; CNPQ::CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL
• Link de acesso: http://repositorio.ufsm.br/handle/1/13558

Awareness of losses caused by diseases in the field is considered one of the most important factors in the integrated management of diseases in Eucalyptus. Erwinia psidii die-back has been reported as one of the major diseases in commercial eucalyptus plantations. The objective of this work was to develop a methodology for the processing and analysis of images obtained by the multispectral sensor in the UAV, so to assist in the characterization of the spectral response of Eucalyptus urugrandis with incidence of the bacterium Erwinia psidii, aiming at discrimination and recognition of levels of disease severity in forest plantations. The assessed settlement is located in the municipality of Minas do Leão - RS, where the flight was carried out with a UAV equipped with a multispectral camera. In the field, the assessment of a portion of disease monitoring was carried out, in which level 1) without occurrence of the disease; 2) small occurrence of the disease: less than 30% of the tree; 3) average occurrence of the disease: up to 50% of the tree; 4) high occurrence of the disease: over 50% of the tree. The flight resulted in two images, one RGB and one multispectral, from which the simulated spectral behavior in the different sensor bands was extracted. The treetops of the monitored portion were manually vectored and classified according to the field survey. Eleven vegetation rates were calculated for the treetops, compared at different levels of severity, and evaluated using parametric and non-parametric statistical tests. These indicated that the Plant Senescence Reflectance Index (PSRI) is the most adequate to differentiate the severity levels of the disease caused by Erwinia psidii in plantations of Eucalyptus urugrandis. As the severity increases, the PSRI increases in the different bands, except in the IVP band. The simulated spectral behavior showed that the severity level 4 trees have higher reflectance than the other levels in the green, red and RedEge bands of the image Multispectral. The absolutely healthy area of the field represented only 30.38% of its total area, while the area with some degree of disease severity corresponded to 52.14%. The remaining area corresponded to regions with low leaf cover, early transition to exposed soil, exposed soil areas, between lines and planting failures. This work characterizes an innovation in the conventional way of quantifying and detecting the presence of pathogen signals in the forest area. The result found in this work indicates that the great advantage of the use of UAVs in forestry is the increase of the spatial and temporal resolution, allowing the analysis of the individualized treetops, besides its ease of handling and speed in the acquisition of data.