Non-destructive wavelet-based despeckling in SAR images

Yuri S. Bekhtin; Andrey A. Bryantsev; Damiao P. Malebo; Alexey A. Lupachev

doi:10.1117/12.2066972

Translator Disclaimer

21 October 2014 Non-destructive wavelet-based despeckling in SAR images

Yuri S. Bekhtin, Andrey A. Bryantsev, Damiao P. Malebo, Alexey A. Lupachev

Author Affiliations +

Proceedings Volume 9243, SAR Image Analysis, Modeling, and Techniques XIV; 92430B (2014) https://doi.org/10.1117/12.2066972
Event: SPIE Remote Sensing, 2014, Amsterdam, Netherlands

Abstract

The suggested wavelet-based despeckling method for multi-look SAR images does not use any thresholding and window processing to avoid ringing artifacts, blurring, fusion of edges, etc. Instead, the logical operation of comparison is applied to wavelet coefficients which are presented in spatial oriented trees (SOTs) of wavelet decomposition calculated for one and the same region of the earth surface during SAR spacecraft flight. Fusion of SAR images is decided by keeping the smallest wavelet coefficients from different SOTs in high frequency subbands (details). The wavelet coefficients related to the low frequency subband (approximation) are processed by another special logical operation providing with a good smoothing. It is because the described procedure depends on properties of the chosen wavelet basis then the library of wavelet bases is applied. The procedure is repeated for each wavelet basis. To select the best SOTs (and hence, the best wavelet basis) there is the special cost function which considers the SOTs as so-called coherent structures and shows which of wavelet bases brings the maximum entropy. The results of computer modeling and comparison with few well-known despeckling procedures have shown the superb quality of the proposed method in the sense of different criteria as PSNR, SSIM, etc.

Citation Download Citation

Yuri S. Bekhtin, Andrey A. Bryantsev, Damiao P. Malebo, and Alexey A. Lupachev "Non-destructive wavelet-based despeckling in SAR images", Proc. SPIE 9243, SAR Image Analysis, Modeling, and Techniques XIV, 92430B (21 October 2014); https://doi.org/10.1117/12.2066972

ACCESS THE FULL ARTICLE