Robust underwater image enhancement method based on natural light and reflectivity.

Area of Science:

• Computer Vision
• Image Processing
• Optical Engineering

Background:

• Underwater image visibility is degraded by scattering, absorption, and unfavorable lighting conditions.
• Existing methods often fail to robustly handle natural and artificial light variations.
• Poor visibility limits applications in marine research, exploration, and surveillance.

Purpose of the Study:

• To develop a robust underwater image enhancement method.
• To improve image quality by accurately estimating natural light and reflectivity.
• To overcome limitations of existing methods in handling diverse lighting scenarios.

Main Methods:

• A novel method employing natural light and reflectivity for underwater image enhancement.
• A dehazing process based on the non-correlation of foreground scene and background light to estimate reflectivity.
• Incorporating the dehazed image into the Lambertian model to remove artificial light and compensating natural light loss using an attenuation ratio map.

Main Results:

• The proposed method effectively addresses scattering and absorption effects.
• Accurate estimation of reflectivity and derivation of least-attenuated natural light.
• Demonstrated superior performance in producing pleasing underwater images across various conditions.

Conclusions:

• The novel method significantly enhances underwater image quality and robustness.
• It provides a more effective solution for visibility issues compared to classical approaches.
• The technique offers improved visual fidelity for underwater imaging applications.