SAR image despeckling using adaptive multiscale products thresholding in directionlet domain

• Published in: Electronics letters Vol. 49; no. 18; pp. 1183 - 1184
• Main Authors: Sethunadh, R, Thomas, T
• Format: Journal Article
• Language: English
• Published: Stevenage The Institution of Engineering and Technology 29.08.2013; Institution of Engineering and Technology
• Subjects: adaptive multiscale products thresholding; adjacent scale coefficients; Applied sciences; Artificial intelligence; Computer science; control theory; systems; Detection, estimation, filtering, equalization, prediction; directional vanishing moments; directionlet domain; directionlet transform; DT multiscale correlation; Earth, ocean, space; edge detection; edge preservation; edge structure enhancement; Exact sciences and technology; External geophysics; GCV technique; generalised cross‐validation technique; image denoising; image enhancement; image reconstruction; Information, signal and communications theory; optimal subband adaptive threshold; Pattern recognition. Digital image processing. Computational geometry; perfect reconstruction transform; radar imaging; Radar, sonar and navigation; Remote sensing, photointerpretation; SAR image despeckling; Signal and communications theory; Signal, noise; speckle; speckle reduction; synthetic aperture radar; synthetic aperture radar image despeckling; Telecommunications and information theory; threshold applied DT coefficients; transforms; weakening noise; speckle; adjacent scale coefficients; transforms; directional vanishing moments; image enhancement; adaptive multiscale products thresholding; radar imaging; perfect reconstruction transform; GCV technique; edge detection; generalised cross-validation technique; directionlet transform; directionlet domain; DT multiscale correlation; synthetic aperture radar image despeckling; edge structure enhancement; weakening noise; threshold applied DT coefficients; edge preservation; image reconstruction; image denoising; synthetic aperture radar; optimal subband adaptive threshold; SAR image despeckling; speckle reduction; microwaves; image processing; Subband decomposition; Noise reduction; Speckle; remote sensing; Cross validation; anisotropic materials; Adaptive method; Noisy image; preservation; Image synthesis; correlation; Multiscale method; Sparse representation; Synthetic aperture radar; Threshold
• DOI: 10.1049/el.2013.2186
• ISSN: 0013-5194; 1350-911X

A directionlet transform (DT)-based multiscale products thresholding is presented which uses a generalised cross-validation (GCV) technique for synthetic aperture radar image despeckling. DT has gained popularity over the past few years as an anisotropic, critically sampled and perfect reconstruction transform with directional vanishing moments along any two directions. In this reported work, the multiscale correlation of DT is exploited by multiplying the adjacent scale coefficients to enhance edge structures while weakening noise. An optimal subband adaptive threshold based on GCV is then computed using this multiscale product. The despeckled image is finally synthesised by using the threshold applied DT coefficients. The employment of DT results in the significant features in images evolving with high magnitudes across scales, while the noise decays rapidly. The proposed scheme outperforms many of the traditional despeckling schemes in terms of speckle reduction and edge preservation.