ADOM+: Accelerating Stripe Noise Removal in Remote Sensing Images With An Advanced ADMM Approach

Abstract: Striped noise in remote sensing image (RSI) often caused by sensor companies significantly impedes the effectiveness of critical applications such as environmental surveillance and disaster responses. To improve this, we propose an improved fracture approach based on the Multiplier Directional Method (ADMM). The extended model includes a robust weight mechanism, evidence-formed initialization strategies, and adaptive impulse-based step sizes to improve both accuracy and efficiency. This model has been rigorously tested on both synthetic data records (such as Cuprite and Pavia University) and real data sets (including Hyperion, M3, Aqua Modis and Terra Modis). This performance is evaluated based on a comprehensive statement of valuation metrics including PSNR, SSIM, D, ICV, and 1 billion. The results consistently show that our approach surpasses major disruption technologies such as WTAF, UTV, GSR, DLS, RBS, LRHP, and GDF. Surprisingly, this method offers a profit of 2.5 dB in PSNR, reducing execution time by about 1x. 15%, even under different strip noise. Keywords: Remote sensing, stripe noise removal, ADMM, optimization, weight-based detection, image restoration.