标题：Multi-temporal object-oriented classifications and analysis of Quickbird scenes at a metropolitan area in Brazil (Belo Horizonte, Minas Gerais State)
期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2006
卷号：XXXVI-4/C42
出版社：Copernicus Publications
摘要：Due to the strong urbanization process occurring in Brazil and of population increase in the cities and to serious impacts on environment, there is a growing need to develop technologies and methodologies which allow to monitor, analyze and take decisions (planning) referring to urban expansion. This paper presents the results of a larger and detailed study done in two quarters of Belo Horizonte, the Minas Gerais State Capital, using high-resolution images from two dates. Automatic land use classification algorithms were applied, using a new paradigm: object-oriented classification. It uses context relations, hierarchy and fuzzy logic that are new and efficient procedures to map urban land cover, specially when analyzing high-resolution images. The analysis of the datasets and ancillary data allowed spatial inferences. Finally an analysis on the quality of the classification is done using Kappa statistics. The class stability was also analyzed
