Implementation of 3D discrete wavelet scheme for space-borne imagery classification and its application

A three dimensional discrete wavelet transform (3D DWT) expands planes of the 2D DWT into volume data. The 3D DWT scheme has advantages of analyzing spectral characteristics in the order of frequency and analyzing changes of spatial information and spectral information simultaneously. Nonetheless, few researchers have attempted to classify multi-spectral images using the 3D DWT. This study aims to apply the 3D DWT to the classification of multi-spectral images and synthetic aperture radar image. To classify these images, we employ two numerical values: The 3D wavelet coefficients and the energy of each band. And then we evaluate their results quantitatively with those of traditional classification techniques including the 2D wavelet scheme. Therefore, the results show that the classification technique of the 3D DWT is more effective than that of traditional techniques, especially in complicated imagery. The accuracy of the 3D energy in SAR imagery is higher. This study provides new numerical values to classify images effectively. Furthermore, these values can be extended to the image retrieval and pattern recognition in high-resolution imagery and this scheme can be employed in image dataset obtained at other times.

[1]  Stéphane Mallat,et al.  A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Feng Wu,et al.  Advanced motion threading for 3D wavelet video coding , 2004, Signal Process. Image Commun..

[3]  G. A Theory for Multiresolution Signal Decomposition : The Wavelet Representation , 2004 .

[4]  T.P.H. Steeghs,et al.  WAVELET BASED VOLUME ATTRIBUTES FOR SEISMIC INTERPRETATION , 2001 .

[5]  Feng Wu,et al.  Studies on spatial scalable frameworks for motion aligned 3D wavelet video coding , 2005, Visual Communications and Image Processing.

[6]  Antonio Ortega,et al.  Efficient wavelet-based predictive Slepian-Wolf coding for hyperspectral imagery , 2006, Signal Process..

[7]  Miriam Scadeng,et al.  Superiority of 3D wavelet-packet denoising in MR microscopy. , 2003, Magnetic resonance imaging.

[8]  Ruola Ning,et al.  Breast volume denoising and noise characterization by 3D wavelet transform. , 2004, Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society.

[9]  Michel Barlaud,et al.  Image coding using wavelet transform , 1992, IEEE Trans. Image Process..

[10]  Joachim M. Buhmann,et al.  Region-based motion compensated 3D-wavelet transform coding of video , 1997, Proceedings of International Conference on Image Processing.