Underwater image enhancement method based on adaptive attenuation-curve prior.

Area of Science:

• Computer Vision
• Image Processing
• Optical Engineering

Background:

• Underwater environments present significant challenges for image acquisition due to light attenuation (absorption and scattering).
• This attenuation causes color degradation and loss of fine details, hindering accurate visual analysis and interpretation.
• Existing underwater image enhancement methods often struggle to effectively address non-uniform attenuation and preserve image fidelity.

Purpose of the Study:

• To develop an advanced underwater image enhancement method.
• To improve color restoration and detail preservation in underwater images.
• To address the limitations of current methods in handling complex underwater optical conditions.

Main Methods:

• Utilized an adaptive attenuation-curve prior for image restoration.
• Employed color channel transfer (CCT) for preprocessing and wavelength-dependent attenuation estimation.
• Incorporated light smoothing, water light estimation, and relative transmission refinement.
• Applied white balance fusion with globally guided image filtering (G-GIF) for color enhancement and edge preservation.

Main Results:

• The proposed method effectively restored color balance and reduced haze in underwater images.
• Significant improvements in edge detail clarity were observed compared to existing techniques.
• The adaptive attenuation-curve prior demonstrated superior performance in handling non-uniform attenuation.

Conclusions:

• The developed underwater image enhancement method offers superior performance in color restoration and de-hazing.
• The integration of adaptive attenuation curves and G-GIF technology effectively preserves image details.
• This approach provides a robust solution for improving the quality of underwater imagery.