Denoising and analysis of synthetic aperture radar images using improved weight threshold technique in curvelet transform frequency domain
DOI: https://doi.org/10.1007/s11042-024-19304-7
IF: 2.577
2024-05-14
Multimedia Tools and Applications
Abstract:Synthetic Aperture Radar (SAR) is a critical instrument in remote sensing, assuming images with changing resolutions depending on weather conditions. However, SAR images often suffer from speckle noise, impacting the quality and interpretability of the derived features. Hence introduced an Improved Curvelet Thresholding technique to proficiently denoise SAR images while retaining important features. The technique uses the Curvelet Transform to analyze SAR images with speckle noise. It uses directional filtering to gather directional information and further filtering at each scale. Curvelet coefficients are derived from these sub-bands for denoising. The median absolute deviation (MAD) is used to estimate the noise level around each coefficient. The Improved Weight Thresholding technique is used to calculate thresholds, with weight shrinkage applied if the coefficient is below the threshold. Following thresholding, the inverse Curvelet transform was employed to reconstruct the image, resulting in a denoised SAR image that effectively preserves edges. Experimental results demonstrate the efficacy of the Improved Curvelet Thresholding technique in reducing speckle noise, etc. when compared to existing techniques. The proposed technique improves overall image quality while successfully reducing noise and suppressing speckle noise. As the result achieved NMV as 59.58, MSD as 2693.9, PSNR as 42d, ENL as 28.32, NSD as 7.176, SSI as 0.0278, UIQI as 0.99 and NV as 0.09 make it an attractive solution for high-quality picture denoising in a variety of applications.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering