Underwater image restoration by structured light and flood light imaging

Area of Science:

• Oceanic engineering
• Optical imaging
• Image processing

Background:

• Underwater optical imaging systems face challenges due to poor visibility caused by light scattering and absorption.
• Degradation of underwater images significantly impacts oceanic engineering tasks.
• Effective removal of backscattered light is crucial for improving underwater image visibility, especially in turbid environments.

Purpose of the Study:

• To develop and present an approach for underwater image recovery.
• To address the issue of poor visibility in underwater optical imaging.
• To completely offset the backscatter component in underwater images.

Main Methods:

• A combined imaging model using structured light imaging and flood light imaging.
• Acquiring a convolutional image via structured light scanning.
• Developing an algorithm to refine the matching of backscatter components between convolutional and flood light images.
• Subtracting the matched backscatter components to create a backscatter-free imaging model.
• Employing a deconvolution process for final image restoration.

Main Results:

• The proposed method successfully offsets the backscatter component in the combined imaging model.
• Image restoration is achieved through deconvolution after backscatter removal.
• Experimental results demonstrate the effectiveness of the approach in underwater environments.

Conclusions:

• The presented approach effectively recovers underwater images by addressing light scattering and absorption.
• Combining structured and flood light imaging offers a robust solution for enhancing visibility in turbid waters.
• The method provides a significant improvement for various oceanic engineering applications requiring clear underwater visuals.