ROUGH WAVELET GRANULAR SPACE AND CLASSIFICATION OF MULTISPECTRAL REMOTE SENSING IMAGE

摘要

properly selected wavelet base and the decomposition level (s) having a shift invariant wavelet transformer form is the use of rough granules wavelet generating a wavelet granules illustration of the feature space for the multi- spectral image , such as a remote sensing image characteristics to build . Relations of the granules through the use of federated time characteristic spatial and / or frequency domain are analyzed individually or in combination . Rough neighboring sets (NRS) is used to select a subset of the features that granular federated further analyze the local and / or related information from neighboring granules . ;