Técnicas de sensoriamento remoto na detecção do Mofo-Branco e da Mancha-Angular do feijoeiro

• Ano de defesa: 2012
• Autor(a) principal: Boechat, Lorena Temponi
• Orientador(a): Pinto, Francisco de Assis de Carvalho
• Banca de defesa: Khoury Junior, Joseph Kalil , Teixeira, Hudson , Santos, Fábio Lúcio
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Doutorado em Engenharia Agrícola
• Departamento: Construções rurais e ambiência; Energia na agricultura; Mecanização agrícola; Processamento de produ
• País: BR
• Palavras-chave em Português: Sensoriamento remoto; Mofo branco; Mancha-angular do feijoeiro
• Palavras-chave em Inglês: Remote sensing; Mold-White; Stain-Angle bean
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA
• Link de acesso: http://locus.ufv.br/handle/123456789/731

This study was conducted with the aim of using remote sensing techniques to detect fungal diseases of common bean: white mold and angular leaf spot. Four experiments were conducted with different doses of fungicides to favor the occurrence of different levels of disease severity. In the four experiments were used a randomized complete block design with different replicates and treatments (fungicides and different doses plus the control). In Experiment 1 and 4 was used the cultivar Madrepérola ; in Experiment 2, the cultivar BRSMG Majestoso ; and in Experiment 3, the Ouro Vermelho cultivar. For measurements of the spectral reflectance factors, it was used a system of dual-fiber, with two spectroradiometers, obtaining data at wavelengths between 400 nm to 850 nm, covering the visible and near infrared bands. Images were acquired of the bean canopy at 3.0 m height on the centerlines of the plot. It was used an imaging system consists of two cameras (one color and one monochrome), they grabbed two images simultaneously of the same scene. In the monochrome camera was attached an optical filter that blocks the visible spectral range for near-infrared (NIR) image acquisition. Thus it was covered the four interested bands: R, G and B with one of the cameras and NIR with the other. In Experiments 2, 3 and 4 digital numeric value images were converted to reflectance values images based on three calibration equations (one for each spectral band of interest: green, red, and near-infrared). It was tested a method for radiometric calibration of the digital numeric values of the images using panels with different gray levels. Estimates of chlorophyll content and leaf area index (LAI) in each plot were measured indirectly by using a portable chlorophyll meter SPAD and a ceptometer, respectively, both measured simultaneously with the reflectance factor acquisition and the images. Graphics were generated with the spectroradiometer values: mean canopy reflectance versus wavelength for each treatment and on each time of data collection, in order to display the spectral progress of disease severity. The values of reflectance factors were used to calculate vegetation indices: NDVI, GNDVI, SAVI, OSAVI, and MCARI1 MCARI2. It was performed a correlation analysis of vegetation indices with disease severity and yield for each date. It also was performed a correlation analysis of SPAD values and LAI with disease severity and yield. The remote sensing techniques can be used for detection of white mold and angular leaf spot in common bean. It was possible to study the spectral behavior of the common bean crop with severity different levels of white mold and angular leaf spot. Assuming a significance level of 10% at 71 DAE, the four experiments found a significant correlation with NDVI and severity of disease. The tested method of radiometric calibration was done, but it was not possible to discriminate between the calibrated and no calibrated image. The results of the correlations between vegetation indices and grain yield are also promising, because in experiments 1 and 2, at 71 DAE, significant correlation was found between the GNDVI and bean yield, and, at 91 DAE, all indices had significant correlation with yield.