Title :
Subband decompositions for hyperspectral image analysis
Author :
Hong, Paul S. ; Smith, Mark J T
Author_Institution :
Center for Signal & Image Process., Georgia Inst. of Technol., Atlanta, GA, USA
Abstract :
Several multidimensional filter banks are described and analyzed for the purpose of processing hyperspectral data. A new octave band directional filter bank (OBDFB) is introduced that is able to isolate directional bands in three-dimensional Fourier space. The new OBDFB is a maximally decimated exactly reconstructing representation with high computational efficiency. The OBDFB is compared with traditional octave band and uniform band decompositions with respect to compaction and feature identification for analysis.
Keywords :
geophysical signal processing; image classification; image reconstruction; image representation; image segmentation; image texture; multidimensional digital filters; remote sensing; OBDFB; compaction; feature identification; hyperspectral data processing; hyperspectral image analysis; multidimensional filter banks; octave band directional filter bank; subband decomposition; three-dimensional Fourier space; uniform band decomposition; Filter bank; Frequency; Hypercubes; Hyperspectral imaging; Hyperspectral sensors; Image analysis; Image coding; Image segmentation; Passband; Principal component analysis;
Conference_Titel :
Signals, Systems and Computers, 2002. Conference Record of the Thirty-Sixth Asilomar Conference on
Conference_Location :
Pacific Grove, CA, USA
Print_ISBN :
0-7803-7576-9
DOI :
10.1109/ACSSC.2002.1197287
