ABSTRACT
Most image analysts would agree that, when extracting urban/suburban information from remotely sensed data, it is more important to have high spatial resolution (often finer than 5 m by 5 m) than high spectral resolution (i.e., a large number of spectral bands) (Jensen and Cowen 1999). However, some researchers (Latty and Hoffer 1981; Irons et al. 1985; Green et al. 1993; Muller 1997) have reported that finer spatial resolution image data do not necessarily improve traditional spectral-based image classification. Moreover, the spectral classification approach has been criticized when fine spatial resolution images are used, especially for urban features (Latty and Hoffer 1981; Markham and Townshend 1981; Woodcock and Strahler 1987; Cushnie 1987; Myint 2001). Traditional image classification meth ods, such as the maximum likelihood classifier, use spectral information (pixel values) as a basis to analyse and classify remote sensing images. They become less efficient when complex urban features are analysed (see Mesev, Chapter 9 in this book).
