학술논문

A New DCT-based Multiresolution Method for Simultaneous Denoising and Fusion of SAR Images
Document Type
Conference
Source
2006 9th International Conference on Information Fusion Information Fusion, 2006 9th International Conference on. :1-8 Jul, 2006
Subject
Communication, Networking and Broadcast Technologies
Computing and Processing
Signal Processing and Analysis
Components, Circuits, Devices and Systems
Image resolution
Noise reduction
Discrete wavelet transforms
Discrete cosine transforms
Speckle
Additive white noise
Spatial resolution
Gaussian noise
Image fusion
Image denoising
Data fusion
synthetic aperture radar
SAR
Discrete Cosine Transform
DCT
multiresolution
denoising
speckle
Language
Abstract
Individual multiresolution techniques for separate image fusion and denoising have been widely researched. We propose a novel multiresolution Discrete Cosine Transform based method for simultaneous image denoising and fusion, demonstrating its efficacy with respect to Discrete Wavelet Transform and Dual-Tree Complex Wavelet Transform. We incorporate the Laplacian pyramid transform multiresolution analysis and a sliding window Discrete Cosine Transform for simultaneous denoising and fusion of the multiresolution coefficients. The impact of image denoising on the results of fusion is demonstrated and advantages of simultaneous denoising and fusion for SAR images are also presented.