Caracterização espectral de cultivares de cafeeiros e monitoramento de parâmetros fitotécnicos após a poda a partir de imagens multiespectrais

• Ano de defesa: 2020
• Autor(a) principal: Freitas, Renato Aurélio Severino de Menezes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Agricultura e Informações Geoespaciais
• Palavras-chave em Português: Cafeeiro; Coffe tree; Decote; Decote-type pruning; Dados espectrais; Spectral data; Índice de vegetação; Vegetation index; ARP; RPA; Sensoriamento remoto; Remote sensing; CNPQ::CIENCIAS AGRARIAS::AGRONOMIA::FITOTECNIA::MANEJO E TRATOS CULTURAIS; Agronomia
• Link de acesso: https://repositorio.ufu.br/handle/123456789/31135; http://doi.org/10.14393/ufu.di.2020.846

The cultivars of Coffea arabica L. in Brazil have different characteristics regarding vegetative development and adaptations to edaphoclimatic conditions. However, due to the interaction between genotype and environment, it is essential to verify the vegetative and productive behavior of genetic materials in each region of cultivation. The objective with this work was to spectrally characterize coffee cultivars and monitor the phytotechnical parameters of arabica coffee plants after pruning. For the spectral characterization of coffee cultivars, biochemical and structural parameters were established at leaf level in order to simulate the vegetation reflectance curve. For the monitoring of phytotechnical parameters, multispectral images were collected using the Mapir Survey 3 camera on board a remotely piloted aircraft (ARP). The data were subjected to statistical analysis, where from the spectral intervals of greatest discrepancy between cultivars, models for estimating agronomic parameters were created. By analyzing the spectral characterization data, it was possible to discriminate the coffee genotypes, and each cultivar presented a different behavior between them, where the largest discrepancies were observed in the visible spectral region. By analyzing data from multispectral images, it was possible to estimate the agronomic parameters of coffee trees after decote-type pruning, being height through the april flight and the near infrared band (Accuracy = 91.87%), crown diameter and length of plagiotropic branches through the april flight and the red band (Accuracy = 89.36% and 82.22%, respectively), number of nodes through the february flight and the near infrared band (Accuracy = 79.48%) and number of plagiotropic branches from the pruning point through the june flight and the near infrared band (Accuracy of the model = 69.57%).