Underwater dynamic polarization imaging without dependence on the background region.

Area of Science:

• Optics
• Image Processing
• Underwater Imaging

Background:

• Active-polarization imaging offers potential for clear underwater vision.
• Previous methods were limited to static targets and required background information for restoration.

Purpose of the Study:

• To develop an underwater dynamic polarization imaging method.
• To overcome limitations of static target and background region dependency in prior research.

Main Methods:

• Image pyramid decomposition and reconstruction technique applied to polarized images.
• Downsampling to create an image pyramid, followed by upsampling for scene reconstruction.
• Method designed to handle moving targets with changing polarization characteristics.

Main Results:

• Effective elimination of backscattered light in underwater images.
• Preservation of crucial target details during image restoration.
• Demonstrated suitability for dynamic targets, achieving practical processing times and superior recovery.

Conclusions:

• The proposed method successfully enables clear underwater vision for dynamic targets.
• It eliminates the need for a background region, broadening applicability.
• Achieves a balance between processing speed and image recovery quality for real-world applications.