HIGH-LEVEL IMPERVIOUS SURFACES CLASSIFICATION IN URBAN ENVIRONMENTS FROM HYPERSPECTRAL IMAGERY

摘要

Remote sensing techniques have great potential in providing accurate and timely information in urban areas. Estimation of impervious surfaces (IS) is rousing widely interests of researchers in monitoring urban development and determining the overall environmental health of a watershed. However, study on IS is complicated due to the complexity of urban infrastructures which include considerable spectral diversity when conduct research at very fine spatial scales. While traditional multi-spectral remote sensing data is inadequate in dealing with high-level IS estimation. With the urgent needs of testing the capability of hyper-spectral imageries in further recognizing different materials of estimated IS, an image equipped with both high spatial and spectral resolution is adopted to demonstrate the whole workflow in using high dimensional data for high-level impervious surface classification, including determining the number of sub-classes and selecting spectrally similar training samples.