Spatial and spectral comparison among IKONOS, CBERS and ASTER images to identify and detect land occupation changes around urban railway in São Paulo - Brazil

The purpose of this article is to present the results of an experiment with the main objective to evaluate the performance and observe practical aspects of principal components and supervised classification algorithms to select attributes, identify and classify urban land use areas around a railway in São Paulo - Brazil, by images with different spatial and spectral resolution. In this experiment were used IKONOS II images of 2002 (R,G, B and NIR channels, geometrically corrected and combined with panchromatic channel), CBERS II images of 2006 (R,G,B and NIR, with 20m of spatial resolution and geometrically corrected) and ASTER images of 2002 (VNIR: channels 1N, 2N, 3N and 3B; SWIR: channels 4 to 9; and TIR: channels 10 to 14, of 15, 30 and 90m de spatial resolution respectively), of one segment of a railway in the metropolitan São Paulo (Brazil) region. The computational procedures were developed using ENVI 4.2 from Research Systems. To conclude, the different sensors can be used by the railway managements as a tool to take decisions related to environmental and social aspects associated to existing land use/occupation and its recently changes identified by the images.