Enhancing Remote Sensing Image Quality with Advanced ADMM-Based Stripe Noise Removal

This study presents an efficient methodology for stripe noise removal in remote sensing images using the Alternating Direction Method of Multipliers (ADMM). The process begins with dataset preparation, where the image matrix is normalized and key characteristics, such as dimensions and peak intensity, are identified. Stripe noise is simulated by introducing vertical streaks with defined intensity ranges, creating a noisy image for testing. The ADMM framework is then employed with carefully initialized parameters, including regularization coefficients and iteration limits, to iteratively separate noise components and reconstruct the underlying image. The algorithm effectively suppresses noise while preserving essential image features, as validated by performance metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and mean absolute error (D). Visual comparisons of clean, noisy, and destriped images highlight the success of the noise removal process. With an execution time of 1.59 seconds, this robust approach enhances remote sensing image quality, enabling more reliable data analysis and supporting applications in environmental monitoring, urban planning, and land classification.