An Improved Frequency Domain Guided Thermal Imager Strips Removal Algorithm Based on LRSID.

Area of Science:

• Earth Observation
• Remote Sensing Technology
• Image Processing

Background:

• Sustainable Development Science Satellite (SDGSAT-1) thermal imaging captures high-resolution ground data.
• Pendulum sweep imaging and detector imperfections cause fringe noise in SDGSAT-1 images.
• Fringe noise degrades the quality and interpretability of thermal remote sensing data.

Purpose of the Study:

• To develop an effective algorithm for removing lateral and vertical fringe noise from SDGSAT-1 thermal images.
• To maintain image detail and clarity while suppressing fringe artifacts.
• To evaluate the proposed algorithm's performance against existing methods.

Main Methods:

• A novel Fringing algorithm based on low-rank-based single-image decomposition (LRSID) was proposed.
• The method involves pretreatment of stripes, processing vertical and horizontal fringes using LLSID, and frequency domain replacement.
• Fourier inverse transformation was applied to reconstruct the final denoised image.

Main Results:

• The proposed LRSID-based algorithm successfully removed simulated and actual fringe noise from SDGSAT-1 thermal images.
• Visual and quantitative comparisons demonstrated superior performance compared to other algorithms.
• The algorithm effectively removed both horizontal and vertical fringes while preserving image details.

Conclusions:

• The developed Fringing algorithm offers an effective solution for noise reduction in SDGSAT-1 thermal imaging.
• This method significantly enhances the quality of thermal remote sensing data for scientific analysis.
• The algorithm's ability to preserve image clarity makes it valuable for various Earth observation applications.