A new HSI denoising method via interpolated block matching 3D and guided filter

Area of Science:

• Remote Sensing
• Image Processing
• Signal Processing

Background:

• Hyperspectral images (HSIs) are susceptible to spatial and spectral noise.
• Effective denoising is crucial for accurate HSI analysis and interpretation.

Purpose of the Study:

• To propose a new hyperspectral image denoising method, IBM3DGF.
• To evaluate the performance of IBM3DGF against state-of-the-art denoising techniques.

Main Methods:

• Inter-spectral correlation analysis to group HSIs.
• Interpolation using adjacent images to enhance resolution.
• Block-Matching and 3D filtering (BM3D) for initial noise reduction.
• Guided image filtering for final denoising.
• Inverse interpolation to reconstruct the HSI.

Main Results:

• The IBM3DGF method demonstrated superior performance in denoising HSIs.
• Evaluations based on spatial and spectral domain noise assessment confirmed effectiveness.
• The method showed significant improvements on both synthetic and real HSIs.

Conclusions:

• The proposed IBM3DGF method is effective in removing spatial and spectral noise from HSIs.
• IBM3DGF offers a promising solution for improving the quality of hyperspectral data.
• This method has potential applications in various fields utilizing HSI data.