Fechar
Metadados

@MastersThesis{Moura:2000:ImMuTM,
               author = "Moura, Valdir",
                title = "Imagens multitemporais TM/Landsat-5 da {\'e}poca de 
                         pr{\'e}-plantio na estimativa de {\'a}reas destinadas a cultura 
                         de ver{\~a}o",
               school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                 year = "2000",
              address = "Sao Jose dos Campos",
                month = "2000-03-21",
             keywords = "AGRONOMIA, ESTATISTICA AGRICOLA, CULTURA DE VERAO, USO DA TERRA, 
                         SISTEMAS DE INFORMACAO GEOGRAFICA, SIG, MAPEADOR TEMATICO 
                         (LANDSAT), LANDSAT 5, CROP INVENTORIES, LAND USE, THEMATIC MAPPERS 
                         (LANDSAT), GEOGRAPHIC INFORMATION SYSTEMS.",
             abstract = "A estimativa das Areas Destinadas a Culturas de Verao (ADCV)e das 
                         Areas Plantadas por Cultura por Municipio (APPCM)estao entre as 
                         principais variaveis envolvidas na previsao de safras usando 
                         sensoriamento remoto. No Brasil, a obtencao destas informacoes e 
                         feita de maneira indireta pelo Instituto Brasileiro de Geografia e 
                         Estatistica (IBGE)e divulgadas atraves do Levantamento Sistematico 
                         de Producao Agricola (LSPA). Devido a dificuldade em adquirir 
                         imagens orbitais no verao, em funcao da grande cobertura de 
                         nuvens, torna-se necessario trabalhar com dados obtidos nas epocas 
                         de pre-plantio, quando se pretende realizar estatisticas agricolas 
                         utilizando sensoriamento remoto. No presente trabalho, avaliou-se 
                         a utilizacao de dados multitemporais do sensor TM/Landsat-5 
                         obtidos nos meses de julho e setembro de 1998 (epoca de 
                         pre-plantio)para a estimativa das ADCVS e APPCM em uma area do 
                         Estado de Sao Paulo composta pelos municipios de Hortolandia, 
                         Sumare, Nova Odessa, Americana, Santa Barbara d'Oeste, Rio das 
                         Pedras, Iracemapolis e Piracicaba. Foram utilizadas tecnicas de 
                         geoprocessamento e processamento digital de imagens para o 
                         tratamento dos dados disponiveis. Utilizaram-se classificacoes 
                         supervisionadas (MAXVER)e nao-supervisionadas (ISOSEG)para os 
                         dados em numero digital, reflectancia e imagens NDVI 
                         multitemporais, na tentativa de discriminar os talhoes de maior 
                         probabilidade de serem cultivados com ADCVS e posteriormente 
                         checadas no campo para a obtencao das APPCMS. A exatidao das 
                         classificacoes foi avaliada atraves de um sistema de amostragem 
                         desenvolvido para a regiao de estudo, e as estimativas de areas 
                         foram corrigidas atraves da matriz de erro construida a partir 
                         destas avaliacoes. As tecnicas de classificacao utilizadas 
                         permitiram uma qualidade classificada como excelente segundo o 
                         indice Kappa. Os resultados obtidos mostraram que a metodologia 
                         aplicada nesta pesquisa e uma alternativa viavel para determinar 
                         com precisao, em nivel municipal, as ADCVS e as APPCM. Porem, 
                         houve diferencas significativas entre estes resultados e as 
                         estimativas divulgadas pelas entidades oficiais. ABSTRACT: The 
                         estimate of the Area of Land to be planted with annual crops and 
                         its stratification by municipality are important variables in crop 
                         forecast. Nowadays in Brazil, these information are obtained by 
                         the Brazilian Institute of Geography and Statistics (IBGE)and 
                         published under the report Assessment of the Agricultural 
                         Production. Due to the high cloud cover in the main crop growing 
                         season (October to March)it is difficult to acquire good orbital 
                         images. Thus, one alternative is to work with remote sensing data 
                         from dates before the crop growing season. This work presents the 
                         use of multitemporal TM/Landsat-5 data gathered on July and 
                         September of 1998 (before the summer growing season)in order to 
                         estimate the area of land to be planted with summer crops in an 
                         area of Sao Paulo State, Brazil. Geographic Information Systems 
                         (GIS)and digital image processing techniques were applied for 
                         treatment of the available data. Supervised and non-supervised 
                         classifications were used for data in digital number and 
                         reflectance formats and for the multitemporal Normalized 
                         Difference Vegetation Index (NDVI)images. The objective was to 
                         discriminate the tracts with higher probability to become planted 
                         with summer crops. Classification accuracy's were evaluated using 
                         a sampling system developed basically for this study region. The 
                         estimated areas were corrected using the error matrix derived from 
                         these evaluations. The classification techniques presented an 
                         excellent level according to the kappa index. The proportion of 
                         crops stratified by municipalities was derived by a field work 
                         during the crop growing season. These proportion coefficients were 
                         applied onto the area of land to be planted with summer crops 
                         (derived from TM/Landsat-5 data). Thus, it was possible to derive 
                         the area of each summer crop by municipality. The discrepancies 
                         between official statistics and our results were attributed to the 
                         sampling and to the stratification procedures. Nevertheless, this 
                         methodology can be improved in order to provide good crop area 
                         estimatives using remote sensing data, despite the cloud cover 
                         during the growing season.",
            committee = "Formaggio, Antonio Roberto (orientador/presidente) and Epiphanio, 
                         Jos{\'e} Carlos Neves (orientador) and Novo, Evlyn M{\'a}rcia de 
                         Moraes and Antunes, Mauro Antonio Homem and Rocha, Jansle Vieira 
                         da",
         englishtitle = "TM/Landsat-5 data from pre crop season to estimate the area to be 
                         planted with summer crop",
                label = "8839",
             language = "pt",
                pages = "122",
                  ibi = "6qtX3pFwXQZ4PKzA/o5LVN",
                  url = "http://urlib.net/rep/6qtX3pFwXQZ4PKzA/o5LVN",
           targetfile = "publicacao.pdf",
        urlaccessdate = "09 mar. 2021"
}


Fechar