Modelo Linear de Mistura Espectral para Mapeamento dos Estágios de Degradação das Pastagens

• Ano de defesa: 2018
• Autor(a) principal: Viçoso, Laíza Castro Brumano
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Geografia (Pontal)
• Palavras-chave em Português: Sensoriamento remoto; Remote sensing; Mapeamento de pastage; Microrregião de Ituiutaba; MLME (modelo linear de mistura espectral); Mapping of pastures; MLME (linear spectral mixing model); Microregion of Ituiutaba.; CNPQ::CIENCIAS HUMANAS::GEOGRAFIA; CNPQ::CIENCIAS EXATAS E DA TERRA::GEOCIENCIAS::GEOFISICA::SENSORIAMENTO REMOTO
• Link de acesso: https://repositorio.ufu.br/handle/123456789/21314; http://dx.doi.org/10.14393/ufu.di.2018.928

Currently, about 70% of the pastures in Brazil are at some stage of degradation, so that the highest concentration of these areas is present in the agricultural border regions such as the north, northeast and center-west according to EMBRAPA (2014). These data show the lack of public policies aimed at the recovery and ordering of these areas. Pastures are considered degraded when their vigor, productivity and recovery capacity are no longer possible. The quality of pasture demanded by animals, as well as the overcoming of harmful effects such as pests, diseases and invasives, end up favoring the advanced degradation of natural resources due to inadequate management. Thus, this research has the purpose of showing the importance of making pastures more productive and also of contributing to their quality and vigor, without the need to open new pasture areas. This fact avoids the substitution of these areas for other uses, such as agriculture and pasture degradation, in addition to making the areas more productive for livestock, an area of economic and social relevance in the Ituiutaba microregion (MRG). The main objective of this research is to map the stages of degradation of MRG pastures of Ituiutaba / MG, through geoprocessing and remote sensing techniques. To this end, specific objectives included: mapping the MRG pastures of Ituiutaba / MG; to identify the most appropriate techniques for the classification of improved and degraded pastures; to separate the stages of degradation of the pastures of the MRG of Ituiutaba and to quantify them. In order to achieve the proposed objectives, the research methodology included the following steps: a bibliographic review that structured the guiding concepts of this research; the acquisition of OLI / Landsat 8 satellite images in the USGS (United States Geological Survey) database with the atmospheric correction; the application of the linear spectral mixing model (MLME) using the Spring 5.5 software to generate the images fraction soil, shade / water and vegetation; the application of classification algorithm IsoSeg, also in Spring 5.5 to extract information of the components soil and vegetation since these correspond the areas destined to the pastures; field work at MRG de Ituiutaba for data analysis; validation of results; final mapping of Ituiutaba MRG pastures and classification as to their stage of degradation. The results showed that the microregion of Ituiutaba has 207,090 hectares of degraded pastures, corresponding to 51.8% of the total of pastures mapped, which is very similar to those indicated by EMBRAPA for national estimates. Moderate pastures totaled 111,619 ha, equivalent to 28% of the total, while improved pastures occupied 80,651 ha of pastureland, accounting for 20.1%. It is considered that the applied methodology was satisfactory, since the MLME as a technique of enhancement of the target reflectance, allowed that the areas destined to the pastures were better identified, since the images fraction soil and vegetation highlighted the targets allowing the interpretation of these for the generation of mapping as well as mathematical graphs and calculations.