Multi-prior underwater image restoration method via adaptive transmission

Area of Science:

• Computer Vision
• Image Processing
• Marine Optics

Background:

• Underwater images suffer from color cast and low contrast due to light absorption and scattering.
• These degradations hinder accurate target detection and visual analysis.
• Existing methods often struggle with accurate transmission map and background light estimation.

Purpose of the Study:

• To develop an advanced underwater image restoration method.
• To address color distortion and detail blur in underwater imagery.
• To improve the accuracy of transmission map and background light estimation.

Main Methods:

• Proposed a multi-priors underwater image restoration method with adaptive transmission (MUAT).
• Developed a dominant channel transmission calculation method using pixel value differences and channel differences.
• Introduced adaptive attenuation ratio calculation and global white balance for enhanced restoration.

Main Results:

• Achieved accurate transmission map estimation.
• Significantly reduced color distortion and detail blur.
• Demonstrated superior visual quality compared to state-of-the-art methods.

Conclusions:

• The MUAT method effectively restores underwater images.
• Accurate transmission and background light estimation are crucial for high-quality restoration.
• The proposed method offers a significant advancement in underwater image processing.