In this chapter, we have presented a remote sensing image information mining framework, which
explores state-of-the-art data mining and databases technologies to retrieve integrated spectral
and spatial information from remote sensing imagery. We extracted texture features using
statistics of Gabor wavelet coefficients to characterize spatial information, and identified land
18
cover and land use information corresponding to spectral reflection using SVM-based
classification. Feature vectors were clustered and indexed in object-oriented databases with
associated raster data stored in image databases. The effectiveness of the system was measured by
the coverage and novelty.